[{"content":"Writing on robotics, AI and machine learning, data engineering, research methodology, and building technical education in Syria — in Arabic and English.\n","date":"13 July 2026","externalUrl":null,"permalink":"/blog/","section":"Blog","summary":"Writing on robotics, AI and machine learning, data engineering, research methodology, and building technical education in Syria — in Arabic and English.\n","title":"Blog","type":"blog"},{"content":"","date":"13 July 2026","externalUrl":null,"permalink":"/categories/","section":"Categories","summary":"","title":"Categories","type":"categories"},{"content":"Mechatronics and AI/ML engineer, researcher, and technical trainer — based in Aleppo, Syria.\nI build robotics and machine-learning systems, publish research, and train the next generation of Syrian engineers and researchers.\n4 years teaching 500\u0026#43; learners taught 3 merged PRs to OpenCV and OpenDR 40 hours of research training delivered Where would you like to go? # Hiring or collaborating? → Work With Me — projects, skills, CV, and consulting.\nLooking to learn? → Learn — courses, workshops, and free roadmaps in Arabic and English.\nPartner or sponsor? → Ventures — Neurobotics and Boundless.\nFeatured # Merged into OpenCV core — two fixes accepted into one of the world\u0026rsquo;s most-used computer-vision libraries, both on non-ASCII path and text handling.\nScientific Research Camp — 40 training hours across 11 workshops, with SANAD and UNFPA Syria. Camp II is running now with 50 researchers.\nOpen source — 3 merged pull requests into OpenCV and OpenDR, including a Unicode fix in OpenCV core.\nCommon questions # Who is Mulham Fetna? Mulham Fetna is a mechatronics and AI/ML engineer, researcher, and technical trainer based in Aleppo, Syria. He builds robotics and machine-learning systems, has had 3 pull requests merged into OpenCV and OpenDR, and has taught 500+ learners across all programs, 2022–2026. He is the founder of Neurobotics, an engineering and technical-education venture, and Boundless, an academic-services venture partnered with SANAD and UNFPA Syria. What does Mulham Fetna specialize in? Mechatronics and robotics (including ROS and embedded systems), machine learning and computer vision, MLOps and data engineering, and Arabic natural language processing — a rare specialism in the region. He also teaches research methodology and academic publishing. Where is Mulham Fetna based? Aleppo, Syria. He works remotely worldwide and teaches in both Arabic and English. Is Mulham Fetna available for hire or consulting? Yes. He takes engineering work, technical consulting, and one-to-one mentorship, remotely and internationally. Contact him at contact@mulhamfetna.com or through the Work With Me page. What is Mulham Fetna\u0026#39;s ORCID? His ORCID identifier is 0009-0006-4432-798X. His research is also listed on Google Scholar and ResearchGate. Get in touch: contact@mulhamfetna.com\n","date":"13 July 2026","externalUrl":null,"permalink":"/","section":"Mulham Fetna","summary":"Mechatronics and AI/ML engineer, researcher, and technical trainer — based in Aleppo, Syria.\nI build robotics and machine-learning systems, publish research, and train the next generation of Syrian engineers and researchers.\n4 years teaching 500+ learners taught 3 merged PRs to OpenCV and OpenDR 40 hours of research training delivered Where would you like to go? # Hiring or collaborating? → Work With Me — projects, skills, CV, and consulting.\nLooking to learn? → Learn — courses, workshops, and free roadmaps in Arabic and English.\nPartner or sponsor? → Ventures — Neurobotics and Boundless.\nFeatured # Merged into OpenCV core — two fixes accepted into one of the world’s most-used computer-vision libraries, both on non-ASCII path and text handling.\nScientific Research Camp — 40 training hours across 11 workshops, with SANAD and UNFPA Syria. Camp II is running now with 50 researchers.\nOpen source — 3 merged pull requests into OpenCV and OpenDR, including a Unicode fix in OpenCV core.\nCommon questions # Who is Mulham Fetna? Mulham Fetna is a mechatronics and AI/ML engineer, researcher, and technical trainer based in Aleppo, Syria. He builds robotics and machine-learning systems, has had 3 pull requests merged into OpenCV and OpenDR, and has taught 500+ learners across all programs, 2022–2026. He is the founder of Neurobotics, an engineering and technical-education venture, and Boundless, an academic-services venture partnered with SANAD and UNFPA Syria. What does Mulham Fetna specialize in? Mechatronics and robotics (including ROS and embedded systems), machine learning and computer vision, MLOps and data engineering, and Arabic natural language processing — a rare specialism in the region. He also teaches research methodology and academic publishing. Where is Mulham Fetna based? Aleppo, Syria. He works remotely worldwide and teaches in both Arabic and English. Is Mulham Fetna available for hire or consulting? Yes. He takes engineering work, technical consulting, and one-to-one mentorship, remotely and internationally. Contact him at contact@mulhamfetna.com or through the Work With Me page. What is Mulham Fetna's ORCID? His ORCID identifier is 0009-0006-4432-798X. His research is also listed on Google Scholar and ResearchGate. Get in touch: contact@mulhamfetna.com\n","title":"Mulham Fetna","type":"page"},{"content":"","date":"13 July 2026","externalUrl":null,"permalink":"/tags/","section":"Tags","summary":"","title":"Tags","type":"tags"},{"content":"","date":"13 July 2026","externalUrl":null,"permalink":"/tags/%D8%A8%D8%AD%D8%AB-%D8%B9%D9%84%D9%85%D9%8A/","section":"Tags","summary":"","title":"بحث علمي","type":"tags"},{"content":" بناء الإمبراطوريات لا يأتي بالصدفة ولا بضربة حظ.. بل بخطى ثابتة ورؤية واضحة.\nخمس سنوات في الجامعة لم أقضِها كأي طالب عادي. كانت رحلة مبنية على هدف واضح، وعمل مستمر لـ 14 ساعة يومياً، طوال أيام الأسبوع. لم تكن هناك أيام عطلة أو فترات استراحة، لأن متعتي الحقيقية كانت ولا تزال في الأثر الذي أصنعه والنتائج التي أراها على أرض الواقع مع طلابي.\nهل كان لدي خطة محكمة بنسبة 100% لتنفيذ هذا الشيء؟ لا أبداً. لكن كان أمامي هدف لدي الاستعداد للوصول إليه مهما كلفني الأمر، وعزيمة للاستمرار مهما كانت الظروف بل وتسخيرها لصالحي. أنا أعمل على بناء كياني الخاص، وصناعة أثر حقيقي ينبع من المبادرة الذاتية وبناء الإنسان.\nوقتي وطاقتي مسخران بالكامل لبناء ما أؤمن به. لم أعتمد على كيان آخر لأستمد منه مصداقيتي، بل حفرت اسم \u0026ldquo;ملهم\u0026rdquo; بالعمل والجهد. اليوم، عندما أدخل أي مؤتمر أو لقاء، يكفي أن أعرّف عن نفسي بـ: \u0026ldquo;مرحباً، أنا ملهم\u0026rdquo;.\nطبعاً، هذا كله ما كان ليتحقق لولا توفيق الله وتيسيره للأمور، فكل ما قمت به هو السعي المستمر، والنتائج تأتي تباعاً.\nولألخص ما حققته خلال السنوات الماضية بإيجاز شديد:\n💻 في الحيز التقني # دربت بشكل شخصي ما يزيد عن 100 متدرب ومتدربة في علوم البيانات وتعلم الآلة، موزعين على أكثر من 8 دول حول العالم، بمحصلة تجاوزت 200 ساعة تدريبية. جزء كبير منهم شقوا طريقهم وحصلوا على وظائف بعد فترة وجيزة من التدريب، ومن كان موظفاً حصل على ترقية لدمجه هذه التقنيات مع تخصصه الأساسي. (علماً أن 80% من المتدربين كانوا من تخصصات طبية، حيوية، واقتصادية، والبقية من الخلفيات الهندسية وعلوم الحاسوب). 📚 في الحيز الأكاديمي # 1. مع طلاب الدراسات العليا والخريجين:\nأسست فريق \u0026ldquo;باوندلس\u0026rdquo; (Boundless)، والذي دربنا من خلاله أكثر من 150 باحثاً وباحثة في حلب (بأكثر من 100 ساعة تدريبية). أكثر من 30 باحثاً نشروا أوراقهم البحثية الأولى أو شاركوا في مؤتمرات. ما يقارب 20 باحثاً يعملون اليوم على أبحاث جديدة سترى النور قريباً، مع تركيزنا على تكامل الاختصاصات لسد الفجوات البحثية. أشغل اليوم منصب مرشد أكاديمي في فريق الإرشاد لدى المؤسسة السورية الألمانية للبحث العلمي (DSFG). 2. مع اليافعين (منح دراسية و STEM):\nأشرفت على أكثر من 20 طالباً وطالبة للتقدم للمنح الخارجية، 3 منهم حصلوا على قبول نهائي ممول بالكامل في أرقى المؤسسات التعليمية. كنت سفيراً لعدد من المسابقات والأولمبيادات العالمية (علوم الفضاء، الرياضيات لليافعين). عملت كمدرب مع هيئة التميز والإبداع (الفريق الوطني لليافعين لأولمبياد الفيزياء) وأحد المدربين العشرين على مستوى سوريا للهاكاثون البرمجي لليافعين لعامين متتاليين. 🚀 في الحيز المهني والاحترافي # أقمت معسكر بناء الهوية الرقمية الاحترافية، والذي نتج عنه استضافتي في جامعة حمص لإلقاء محاضرة لطلاب كلية الهندسة المعلوماتية. محصلة المستفيدين تجاوزت الـ 300 طالب وطالبة. 💡 في حيز ريادة الأعمال # عملت كمرشد في 3 مسابقات ريادة أعمال سورية (تقني/طبي/بيئي)، ساعدت خلالها الشباب على صقل أفكارهم وبناء نماذج عمل (Business Models) واضحة ومسارات تدفق مالي (Revenue Streams) دقيقة لعرضها على كبار رواد الأعمال والمستثمرين. اليوم، عمري 22 سنة.. دربت أكثر من 600 شخص من 8 دول مختلفة، وحققت ما يزيد عن 500 ساعة تدريبية. وما زلت أرى أن هذه هي البداية فقط. يوماً ما، بإذن الله، سيكون هناك كيان مؤسساتي يحمل اسم \u0026ldquo;ملهم\u0026rdquo; يظلل كل هذه الأنشطة وأنشطة أخرى ما زالت قيد التخطيط. لنتوسع في كل مدينة سورية، ثم الوطن العربي، ثم ننطلق للعالمية.\nومن هنا، أعلن جاهزيتي التامة للتعاون الاحترافي واستضافة محاضرات أو ورش عمل في المدن السورية دون أي مقابل مادي. أقدم هذا العرض لأن هدفي بعيد المدى، ورؤيتي للمستقبل أكبر بكثير من أي عائد مالي قريب.\nتريد استضافتي؟ كل شروط التعاون ونموذج تقديم الطلب موجودة في صفحة استضافة ملهم — هي الطريقة الرسمية لتنسيق أي محاضرة أو ورشة عمل. ","date":"13 July 2026","externalUrl":null,"permalink":"/blog/building-impact/","section":"Blog","summary":" بناء الإمبراطوريات لا يأتي بالصدفة ولا بضربة حظ.. بل بخطى ثابتة ورؤية واضحة.\nخمس سنوات في الجامعة لم أقضِها كأي طالب عادي. كانت رحلة مبنية على هدف واضح، وعمل مستمر لـ 14 ساعة يومياً، طوال أيام الأسبوع. لم تكن هناك أيام عطلة أو فترات استراحة، لأن متعتي الحقيقية كانت ولا تزال في الأثر الذي أصنعه والنتائج التي أراها على أرض الواقع مع طلابي.\nهل كان لدي خطة محكمة بنسبة 100% لتنفيذ هذا الشيء؟ لا أبداً. لكن كان أمامي هدف لدي الاستعداد للوصول إليه مهما كلفني الأمر، وعزيمة للاستمرار مهما كانت الظروف بل وتسخيرها لصالحي. أنا أعمل على بناء كياني الخاص، وصناعة أثر حقيقي ينبع من المبادرة الذاتية وبناء الإنسان.\nوقتي وطاقتي مسخران بالكامل لبناء ما أؤمن به. لم أعتمد على كيان آخر لأستمد منه مصداقيتي، بل حفرت اسم “ملهم” بالعمل والجهد. اليوم، عندما أدخل أي مؤتمر أو لقاء، يكفي أن أعرّف عن نفسي بـ: “مرحباً، أنا ملهم”.\nطبعاً، هذا كله ما كان ليتحقق لولا توفيق الله وتيسيره للأمور، فكل ما قمت به هو السعي المستمر، والنتائج تأتي تباعاً.\nولألخص ما حققته خلال السنوات الماضية بإيجاز شديد:\n💻 في الحيز التقني # دربت بشكل شخصي ما يزيد عن 100 متدرب ومتدربة في علوم البيانات وتعلم الآلة، موزعين على أكثر من 8 دول حول العالم، بمحصلة تجاوزت 200 ساعة تدريبية. جزء كبير منهم شقوا طريقهم وحصلوا على وظائف بعد فترة وجيزة من التدريب، ومن كان موظفاً حصل على ترقية لدمجه هذه التقنيات مع تخصصه الأساسي. (علماً أن 80% من المتدربين كانوا من تخصصات طبية، حيوية، واقتصادية، والبقية من الخلفيات الهندسية وعلوم الحاسوب). 📚 في الحيز الأكاديمي # 1. مع طلاب الدراسات العليا والخريجين:\n","title":"بناء الإمبراطوريات لا يأتي بالصدفة","type":"blog"},{"content":"","date":"13 July 2026","externalUrl":null,"permalink":"/tags/%D8%AA%D8%AF%D8%B1%D9%8A%D8%A8/","section":"Tags","summary":"","title":"تدريب","type":"tags"},{"content":"","date":"13 July 2026","externalUrl":null,"permalink":"/tags/%D8%B1%D8%A4%D9%8A%D8%A9/","section":"Tags","summary":"","title":"رؤية","type":"tags"},{"content":"","date":"13 July 2026","externalUrl":null,"permalink":"/tags/%D8%B1%D9%8A%D8%A7%D8%AF%D8%A9/","section":"Tags","summary":"","title":"ريادة","type":"tags"},{"content":"","date":"13 July 2026","externalUrl":null,"permalink":"/categories/%D9%85%D9%82%D8%A7%D9%84%D8%A7%D8%AA/","section":"Categories","summary":"","title":"مقالات","type":"categories"},{"content":"","date":"7 July 2026","externalUrl":null,"permalink":"/tags/arxiv/","section":"Tags","summary":"","title":"Arxiv","type":"tags"},{"content":"","date":"7 July 2026","externalUrl":null,"permalink":"/tags/citation/","section":"Tags","summary":"","title":"Citation","type":"tags"},{"content":"","date":"7 July 2026","externalUrl":null,"permalink":"/tags/doi/","section":"Tags","summary":"","title":"Doi","type":"tags"},{"content":"","date":"7 July 2026","externalUrl":null,"permalink":"/tags/github/","section":"Tags","summary":"","title":"Github","type":"tags"},{"content":"","date":"7 July 2026","externalUrl":null,"permalink":"/tags/licensing/","section":"Tags","summary":"","title":"Licensing","type":"tags"},{"content":"","date":"7 July 2026","externalUrl":null,"permalink":"/tags/open-science/","section":"Tags","summary":"","title":"Open Science","type":"tags"},{"content":"","date":"7 July 2026","externalUrl":null,"permalink":"/tags/orcid/","section":"Tags","summary":"","title":"ORCID","type":"tags"},{"content":" Publishing Research \u0026amp; Software: The Complete A–Z Guide # When you finish a piece of research or a software tool, the work is only half done. The other half is making it permanent, licensed, discoverable, and citable — so that years from now someone can find exactly what you made, use it under clear terms, and give you credit.\nI recently took a research tool of mine from an untitled notebook to a polished, Apache-licensed, permanently-archived release with a real DOI. This guide is the complete map of that territory: every concept, every option, and — the part most tutorials skip — when and why to choose each one. It\u0026rsquo;s long on purpose. Read it once end-to-end, then keep it as a reference.\nPart 1 — The concepts you must get right first # Most confusion in this space comes from mixing up four different things. Let\u0026rsquo;s separate them.\n1.1 Copyright vs. license (they are not the same) # Copyright is automatic. The instant you write code or text, you own the copyright. You do not need to register anything to \u0026ldquo;own\u0026rdquo; your work. So a license is not about claiming ownership — that\u0026rsquo;s already yours. A license is a grant of permission. It tells other people what they may do with your work and what they must do in return. Here\u0026rsquo;s the counter-intuitive part: if you publish code with no license, nobody may legally use it at all. \u0026ldquo;No license\u0026rdquo; means \u0026ldquo;all rights reserved\u0026rdquo; — the most restrictive state possible. For research you want people to use and cite, that\u0026rsquo;s the opposite of what you need. A license is what opens your work up, on your terms.\n1.2 A DOI (Digital Object Identifier) # A DOI looks like 10.5281/zenodo.21236404. It is a permanent address for a specific digital object, resolvable by prefixing https://doi.org/.\nA normal URL breaks when a site moves — this is called link rot. A DOI is a promise, maintained by a global registration agency, that the identifier will always resolve to your object, even if the file physically moves. That permanence is why journals, theses, and grant reports require DOIs: a reviewer in ten years must still be able to find precisely what you cited.\n1.3 Persistent identifiers, in general # DOIs are one kind of persistent identifier (PID). The ones a researcher meets:\nIdentifier Identifies Issued by DOI an output (paper, dataset, software release) DataCite / Crossref ORCID you, the researcher (a 16-digit iD) ORCID.org SWHID a snapshot of source code Software Heritage ROR a research organization ROR.org You want your outputs (DOIs) linked to you (ORCID). More on ORCID in Part 6.\n1.4 The mental model that makes it all click # Thing Role Analogy GitHub / GitLab where the living, changing work happens (mutable) your writing desk Zenodo / figshare a permanent archive storing a frozen copy (immutable) a national library DOI a permanent address that always resolves to the work the library ISBN + catalog card GitHub can change or be deleted; a DOI\u0026rsquo;d archive record cannot. That permanence, plus a fixed date, is what turns \u0026ldquo;a repo\u0026rdquo; into \u0026ldquo;a citable, provable research artifact.\u0026rdquo;\nPart 2 — Licensing, in full # This is where people freeze up. Let\u0026rsquo;s make it a decision, not a mystery.\n2.1 Why license at all — the four goals # Enable legal use. Without a license, others literally cannot use your work. Require attribution. Every open license makes credit a condition of use — this is your single most practical anti-plagiarism tool. Set boundaries. Commercial use allowed? Patents granted? Must derivatives stay open? Disclaim liability. The \u0026ldquo;AS IS, no warranty\u0026rdquo; clause protects you if someone\u0026rsquo;s use of your code causes harm. 2.2 The three families # Permissive — \u0026ldquo;do almost anything, just keep my name.\u0026rdquo;\nLicense Key trait Best for MIT shortest; attribution only simple projects, maximum adoption BSD-2 / BSD-3 like MIT; 3-clause adds a no-endorsement rule academic code Apache-2.0 attribution + explicit patent grant + NOTICE file most research software; anything patent-adjacent Copyleft — \u0026ldquo;you may use it, but derivatives must stay open.\u0026rdquo;\nLicense Key trait Best for LGPL weak copyleft; can be linked by closed software libraries you want reused but kept open MPL-2.0 file-level copyleft (weak) middle ground GPL-3.0 strong; distributing a derivative forces GPL force downstream openness AGPL-3.0 strongest; even running it as a web service triggers the openness requirement SaaS-proof; maximum control Content / data — for things that aren\u0026rsquo;t code.\nLicense Key trait Best for CC BY 4.0 reuse with attribution papers, figures, docs CC BY-SA attribution + share-alike wikis, remixable content CC BY-NC non-commercial only when you want to block commercial reuse CC0 public-domain dedication (no rights kept) datasets you want maximally reusable ⚠️ Do not use Creative Commons for source code, and don\u0026rsquo;t use software licenses for data. CC licenses don\u0026rsquo;t address patents or source-vs-binary; software licenses don\u0026rsquo;t fit prose/data.\n2.3 How to choose — a quick decision guide # \u0026ldquo;I want it used as widely as possible, including by companies\u0026rdquo; → MIT or Apache-2.0 (Apache if patents might ever matter). Best for citations and adoption. \u0026ldquo;Anyone who improves it must share back\u0026rdquo; → GPL-3.0. \u0026quot;…even if they only run it as a hosted service\u0026quot; → AGPL-3.0. Strictest; note it deters corporate adoption, which can reduce reuse and citations. \u0026ldquo;It\u0026rsquo;s a library meant to be embedded\u0026rdquo; → LGPL or a permissive license. \u0026ldquo;It\u0026rsquo;s data, figures, or a paper\u0026rdquo; → a CC license (or CC0 for maximum data reuse). There\u0026rsquo;s a real trade-off: permissive = maximum reuse/citations; copyleft = maximum control. Pick based on which you value more for this project.\n2.4 How to actually apply a license # Add a LICENSE file with the full text. Grab canonical text from choosealicense.com or, on the command line: gh api /licenses/apache-2.0 --jq \u0026#39;.body\u0026#39; \u0026gt; LICENSE Then fill placeholders like [yyyy] and [name of copyright owner]. Use the license\u0026rsquo;s SPDX identifier (e.g. Apache-2.0, MIT, AGPL-3.0-or-later) consistently in your CITATION.cff, package metadata, and archive deposit. Changing a license later? For a solo project it\u0026rsquo;s your call, but once there are outside contributors, relicensing may require their agreement — so choose deliberately up front. Part 3 — CITATION.cff: telling people how to cite you # A CITATION.cff file (Citation File Format — YAML) sits in your repo root. GitHub reads it and renders a \u0026ldquo;Cite this repository\u0026rdquo; button that spits out APA/BibTeX automatically.\nThe questions to answer before writing it: title; type (software/dataset); one-sentence abstract; version and release date; license (SPDX id); authors (name + ORCID each); 4–6 keywords.\ncff-version: 1.2.0 message: \u0026#34;If you use this software, please cite it as below.\u0026#34; title: \u0026#34;Your Project Title\u0026#34; abstract: \u0026gt;- One or two sentences describing what it does. type: software authors: - family-names: Fetna given-names: Mulham orcid: \u0026#34;https://orcid.org/0009-0006-4432-798X\u0026#34; version: 1.0.0 date-released: \u0026#34;2026-07-07\u0026#34; license: Apache-2.0 url: \u0026#34;https://github.com/OWNER/REPO\u0026#34; repository-code: \u0026#34;https://github.com/OWNER/REPO\u0026#34; keywords: [keyword-one, keyword-two, keyword-three] Once you have a DOI (Part 5), add it here too:\ndoi: 10.5281/zenodo.21236404 # concept DOI (all versions) identifiers: - type: doi value: 10.5281/zenodo.21236404 description: Concept DOI (always latest) - type: doi value: 10.5281/zenodo.21236406 description: Version 1.0.0 Part 4 — The publishing landscape: every option, and when to use it # This is the section people most often get wrong: they think it\u0026rsquo;s \u0026ldquo;arXiv or Zenodo.\u0026rdquo; It isn\u0026rsquo;t. Different platforms archive different objects. Here\u0026rsquo;s the whole map.\n4.1 Preprint servers — for the paper (the narrative) # A preprint server hosts your written manuscript (PDF/LaTeX) before (or alongside) peer review. It timestamps your findings and disseminates them fast; moderation is light, not peer review.\nServer Field arXiv physics, CS, math, engineering, quantitative bio bioRxiv / medRxiv biology / medicine ChemRxiv chemistry (relevant to materials work) TechRxiv engineering / IEEE EarthArXiv, SSRN, … earth sciences, social sciences, etc. When: you have a paper to share. Why: establish priority for your results and get feedback early. (arXiv now also issues DOIs for submissions.)\n4.2 Artifact repositories — for software, data, figures # These archive the outputs behind the paper and mint DOIs.\nRepository Strengths Notes Zenodo free, CERN-run, GitHub integration, any file type the default for code + data figshare polished UI, good for figures/posters some features commercial Dryad curated datasets, strong in bio/eco has a publication fee OSF (Open Science Framework) project management + storage + preregistration free; ties a whole project together Harvard Dataverse institutional-grade data hosting great for datasets When: you have code, a dataset, figures, slides, a poster — anything that should be independently citable. Why: permanence + a DOI. This guide\u0026rsquo;s walkthrough uses Zenodo.\n4.3 Software Heritage — the automatic safety net # Software Heritage automatically archives all public source code it can find, giving each snapshot a SWHID. You can also \u0026ldquo;Save code now\u0026rdquo; manually. When: always — it\u0026rsquo;s passive insurance. Why: even if GitHub and you both vanish, the code survives.\n4.4 Software journals — peer-reviewed credit for the code itself # Venue What you get JOSS (Journal of Open Source Software) free; peer-reviews your software; issues a short paper + DOI; very reputable SoftwareX (Elsevier) peer-reviewed software paper JORS (Journal of Open Research Software) open-access software paper When: your software is a genuine research contribution and you want a peer-reviewed, citable paper about it. Why: it converts \u0026ldquo;a repo\u0026rdquo; into \u0026ldquo;a publication\u0026rdquo; on your CV. For most research tools, JOSS is the highest-value, lowest-cost option — strongly recommended.\n4.5 Package registries — distribution, NOT citation # PyPI (pip), npm, conda, CRAN exist so people can install your software. They are not archives and give no DOI. Publish here in addition to Zenodo — never instead of.\n4.6 The infrastructure behind the scenes (so the names stop being mysterious) # DataCite — the DOI registration agency for data and software. Zenodo mints your DOIs through DataCite. Crossref — the DOI agency for journal articles (most published papers). OpenAIRE — Open Access Infrastructure for Research in Europe. An EU-funded initiative that aggregates metadata about publications, data, and software from thousands of repositories into one linked research graph, links outputs to EU funding, and powers discovery. Zenodo was created by CERN under OpenAIRE, so depositing on Zenodo automatically feeds your work into the European open-science graph. Think of OpenAIRE as the umbrella + discovery layer, and Zenodo as one concrete repository inside it. ORCID — the persistent ID for you (Part 6). 4.7 A complete modern release often uses several at once # Manuscript on arXiv/ChemRxiv (or a journal) → code + data on Zenodo (DOI) → peer-reviewed via JOSS → installable from PyPI → passively archived by Software Heritage → everything linked to your ORCID. All cross-referenced by DOI.\nPart 5 — The full walkthrough: GitHub → Zenodo → DOI # This is the exact, repeatable process. Follow it top to bottom.\n5.1 Prepare the repository # Before archiving, make the repo look like a research artifact:\nREADME.md — title, badges, install, usage, and — crucially for integrity — an honest \u0026ldquo;Limitations\u0026rdquo; section if your model/tool has them. LICENSE — chosen per Part 2. CITATION.cff — per Part 3. requirements.txt / manifest, .gitignore, optionally CONTRIBUTING.md. A .zenodo.json (optional but recommended) to control the deposit\u0026rsquo;s metadata: { \u0026#34;title\u0026#34;: \u0026#34;Your Project Title\u0026#34;, \u0026#34;description\u0026#34;: \u0026#34;One-sentence abstract.\u0026#34;, \u0026#34;upload_type\u0026#34;: \u0026#34;software\u0026#34;, \u0026#34;access_right\u0026#34;: \u0026#34;open\u0026#34;, \u0026#34;license\u0026#34;: \u0026#34;Apache-2.0\u0026#34;, \u0026#34;creators\u0026#34;: [{ \u0026#34;name\u0026#34;: \u0026#34;Fetna, Mulham\u0026#34;, \u0026#34;orcid\u0026#34;: \u0026#34;0009-0006-4432-798X\u0026#34; }], \u0026#34;keywords\u0026#34;: [\u0026#34;keyword-one\u0026#34;, \u0026#34;keyword-two\u0026#34;] } Push it to GitHub:\ngit init -b main \u0026amp;\u0026amp; git add -A \u0026amp;\u0026amp; git commit -m \u0026#34;Initial release\u0026#34; gh repo create REPO --public --source=. --remote=origin --description \u0026#34;…\u0026#34; --push gh repo edit --add-topic topic-one,topic-two 5.2 One-time Zenodo setup # Go to zenodo.org → Log in with GitHub → authorize. Connect your ORCID in Zenodo profile settings (so deposits link to you). 5.3 Enable the repository — and mind the order # Open zenodo.org/account/settings/github/ → find the repo → flip its switch ON (click Sync now if it isn\u0026rsquo;t listed). This installs a webhook on the repo. ⚠️ The #1 mistake — ordering. You must enable the repo in Zenodo before publishing the release. Zenodo only archives releases created after the webhook exists. Publish first and Zenodo never sees it — nothing happens, and you\u0026rsquo;ll think it\u0026rsquo;s broken. This is almost certainly why it \u0026ldquo;didn\u0026rsquo;t work\u0026rdquo; the last time you tried.\n5.4 Publish a GitHub Release — the trigger # On GitHub: Releases → Draft a new release → tag v1.0.0 → write notes → Publish. (Or gh release create v1.0.0 --title \u0026quot;…\u0026quot; --notes \u0026quot;…\u0026quot;.) Only a Release triggers Zenodo — ordinary commits do not. A commit is a daily edit; a Release is you declaring \u0026ldquo;this exact state is a finished, citable version.\u0026rdquo;\n5.5 Collect your DOI # Within ~a minute, the Zenodo GitHub page shows your record. Zenodo mints two DOIs: a concept DOI — always resolves to the latest version (cite this for \u0026ldquo;the software\u0026rdquo;); a version DOI — points to this release forever (cite this for reproducibility). Add a badge to your README. Wire it to the numeric repo id (gh api repos/OWNER/REPO --jq .id) so it needs no DOI number in advance and auto-fills:\n[![DOI](https://zenodo.org/badge/REPO_ID.svg)](https://zenodo.org/badge/latestdoi/REPO_ID) Then write the real DOIs into CITATION.cff and your README\u0026rsquo;s BibTeX (Part 3).\n5.6 Releasing new versions later # Just publish another Release (v1.1.0, …). Zenodo auto-archives it with a new version DOI under the same concept DOI. Bump version and date-released in CITATION.cff first.\n5.7 Work that isn\u0026rsquo;t on GitHub # For a thesis PDF, a raw dataset, a poster: skip GitHub entirely. On Zenodo click New upload, drag the files, fill in metadata, Publish → instant DOI. Zenodo also supports embargoed records — the DOI and date are reserved now, but files stay hidden until a date you choose (perfect for \u0026ldquo;I want proof I did this first, but can\u0026rsquo;t reveal it until my paper is accepted\u0026rdquo;).\nPart 6 — ORCID: a permanent ID for you # An ORCID iD (e.g. 0009-0006-4432-798X) is a free, permanent identifier that disambiguates you from every other researcher with a similar name, forever.\nWhy it matters: names change and collide; ORCID doesn\u0026rsquo;t. Connect it to Zenodo, your GitHub profile, journals, and your CITATION.cff, and every DOI you mint automatically ties into one verifiable record of everything you\u0026rsquo;ve produced. Do this once, early — it compounds over a career. Register at orcid.org.\nPart 7 — Protecting your work from plagiarism # Be clear-eyed: nothing technical prevents someone from copying public files. What you can do is make your authorship provable and dated, and copying legally risky. Layer these:\n1. Establish priority with a timestamp — the big one. A DOI\u0026rsquo;d archive record is third-party-witnessed proof that you had this work on this date. If someone later claims it, you point to a permanent record predating theirs. Your git commit history is a second dated authorship trail. Counter-intuitively, publishing early protects you: a private file proves nothing about when you made it; a dated public record proves everything.\n2. Set enforceable terms with a license. Every open license makes attribution a legal condition. Strip your name and the user is in license violation — grounds for a takedown or a formal complaint. Copyleft (GPL/AGPL) goes further, forcing derivatives to stay open and credited.\n3. Make citation the path of least resistance. CITATION.cff + a visible DOI badge mean the easy, normal thing is to cite you. Most academic \u0026ldquo;plagiarism\u0026rdquo; is lazy non-citation, not malice — remove the excuse.\n4. If it\u0026rsquo;s genuinely sensitive / unpublished. Keep the repo private until you\u0026rsquo;re ready, and use an embargoed Zenodo deposit to lock in the date now while hiding the files until your paper publishes.\n5. Enforcement, if it happens. GitHub has a DMCA takedown process for copied repos; your institution\u0026rsquo;s research-integrity office handles academic misconduct. Your dated DOI + git history are the evidence.\nRecommended posture for work you want credit for: public repo + a license + Zenodo DOI + ORCID. For secret pre-publication work: private repo + embargoed deposit.\nPart 8 — Putting it together: a reusable checklist # ONE-TIME □ Register an ORCID (orcid.org) □ Log in to Zenodo with GitHub; connect your ORCID PER PROJECT □ Write README (with honest limitations), choose + add LICENSE □ Add CITATION.cff and .zenodo.json □ Push to GitHub (public), add topics □ Enable the repo in Zenodo ← BEFORE the release □ Publish a GitHub Release (v1.0.0) → DOI is minted □ Add the DOI badge; write DOIs into CITATION.cff + README □ (Optional) Submit to JOSS for a peer-reviewed software paper □ (Optional) \u0026#34;Save code now\u0026#34; on Software Heritage EACH UPDATE □ Bump version/date in CITATION.cff → publish a new Release → new version DOI Quick reference # I want to… Use Share a paper/preprint arXiv, bioRxiv, ChemRxiv, TechRxiv Make code/data citable with a DOI Zenodo (or figshare, Dryad, OSF) Get a peer-reviewed software paper JOSS, SoftwareX, JORS Let people install my software PyPI, npm, conda (no DOI) Auto-archive my source code Software Heritage Identify myself permanently ORCID Maximize reuse/adoption MIT / Apache-2.0 Force derivatives to stay open GPL-3.0 / AGPL-3.0 License data/figures/prose CC BY / CC0 Conclusion # Publishing well isn\u0026rsquo;t bureaucracy — it\u0026rsquo;s how your work becomes permanent, usable, and credited instead of a file that quietly rots on a hard drive. The core loop is short: license it, archive it, get a DOI, tie it to your ORCID. Do that once and it becomes muscle memory for every project after.\nThis guide is drawn from actually taking one of my own research tools through the entire process. If you get stuck on the Zenodo step, re-read Part 5.3 — the ordering trap catches almost everyone.\n","date":"7 July 2026","externalUrl":null,"permalink":"/tutorials-guides/research-software-publishing-doi-guide/","section":"Tutorials \u0026 Technical Guides","summary":"Everything I learned taking a research tool from a messy script to a licensed, permanently-archived, citable release with a DOI — the concepts, every option, and exactly when and why to use each.","title":"Publishing Research \u0026 Software: The Complete A–Z Guide to Licenses, DOIs, and Making Your Work Citable","type":"tutorials-guides"},{"content":"","date":"7 July 2026","externalUrl":null,"permalink":"/tags/research-software/","section":"Tags","summary":"","title":"Research Software","type":"tags"},{"content":"","date":"7 July 2026","externalUrl":null,"permalink":"/categories/tutorials--guides/","section":"Categories","summary":"","title":"Tutorials \u0026 Guides","type":"categories"},{"content":"A Note on Authenticity and Craftsmanship\nThe tutorials and guides presented in this section represent genuine technical challenges and solutions derived from my professional practice. Each article documents real-world scenarios I have personally encountered, investigated, and resolved—no synthetic examples or theoretical hypotheticals. These are battle-tested implementations, debugging journeys, and architectural decisions drawn from production environments and complex development workflows.\nWhile the underlying experiences and technical solutions remain entirely my own, the published versions have undergone careful refinement with the assistance of large language models. This editorial process serves a singular purpose: transforming my initial raw documentation—often hastily scribbled during incident response or deep in exploratory debugging sessions—into accessible, well-structured narratives that prioritize clarity and reader comprehension. The technical accuracy, insights, and methodologies remain authentically mine; only the presentation has been polished to ensure these lessons are communicated with the precision and readability they deserve.\nConsider this collection a curated archive of professional problem-solving, rendered in its most digestible form without sacrificing the authenticity of hard-won experience.\n","date":"7 July 2026","externalUrl":null,"permalink":"/tutorials-guides/","section":"Tutorials \u0026 Technical Guides","summary":"A Note on Authenticity and Craftsmanship\nThe tutorials and guides presented in this section represent genuine technical challenges and solutions derived from my professional practice. Each article documents real-world scenarios I have personally encountered, investigated, and resolved—no synthetic examples or theoretical hypotheticals. These are battle-tested implementations, debugging journeys, and architectural decisions drawn from production environments and complex development workflows.\nWhile the underlying experiences and technical solutions remain entirely my own, the published versions have undergone careful refinement with the assistance of large language models. This editorial process serves a singular purpose: transforming my initial raw documentation—often hastily scribbled during incident response or deep in exploratory debugging sessions—into accessible, well-structured narratives that prioritize clarity and reader comprehension. The technical accuracy, insights, and methodologies remain authentically mine; only the presentation has been polished to ensure these lessons are communicated with the precision and readability they deserve.\nConsider this collection a curated archive of professional problem-solving, rendered in its most digestible form without sacrificing the authenticity of hard-won experience.\n","title":"Tutorials \u0026 Technical Guides","type":"tutorials-guides"},{"content":"","date":"7 July 2026","externalUrl":null,"permalink":"/tags/zenodo/","section":"Tags","summary":"","title":"Zenodo","type":"tags"},{"content":"","date":"13 April 2026","externalUrl":null,"permalink":"/tags/google-scholar/","section":"Tags","summary":"","title":"Google Scholar","type":"tags"},{"content":"","date":"13 April 2026","externalUrl":null,"permalink":"/tags/researchgate/","section":"Tags","summary":"","title":"ResearchGate","type":"tags"},{"content":"","date":"13 April 2026","externalUrl":null,"permalink":"/tags/scopus/","section":"Tags","summary":"","title":"Scopus","type":"tags"},{"content":"","date":"13 April 2026","externalUrl":null,"permalink":"/tags/student-email-syria/","section":"Tags","summary":"","title":"Student Email Syria","type":"tags"},{"content":"","date":"13 April 2026","externalUrl":null,"permalink":"/categories/tutorials-guides/","section":"Categories","summary":"","title":"Tutorials-Guides","type":"categories"},{"content":"","date":"13 April 2026","externalUrl":null,"permalink":"/tags/university-email-syria/","section":"Tags","summary":"","title":"University Email Syria","type":"tags"},{"content":"","date":"13 April 2026","externalUrl":null,"permalink":"/tags/%D8%A7%D8%B3%D8%AA%D8%AE%D8%B1%D8%A7%D8%AC-%D8%A7%D9%8A%D9%85%D9%8A%D9%84-%D8%AC%D8%A7%D9%85%D8%B9%D9%8A/","section":"Tags","summary":"","title":"استخراج ايميل جامعي","type":"tags"},{"content":"","date":"13 April 2026","externalUrl":null,"permalink":"/tags/%D8%A7%D9%8A%D9%85%D9%8A%D9%84-%D8%AC%D8%A7%D9%85%D8%B9%D9%8A/","section":"Tags","summary":"","title":"ايميل جامعي","type":"tags"},{"content":" دليل الإيميل الجامعي في سوريا 2026: الاستخراج، الفوائد، والقيود # آخر تحديث ومراجعة: 2026-04-13\nالمجال: الجامعات السورية (الإجراءات تختلف حسب الجامعة).\nللتوثيق والمنهج التحريري: Authenticity \u0026amp; Editorial Policy + About.\nاختر المسار المناسب مباشرة # إذا هدفك استخراج الإيميل بسرعة:\nكيفية استخراج الإيميل الجامعي في الجامعات السورية (خطوة بخطوة) إذا هدفك تعرف شو بيفتح الإيميل وشو بيبقى مقفول:\nفوائد الإيميل الجامعي في سوريا وحدوده: المسموح وغير المتاح لمن هذا الدليل؟ وماذا لا يغطي؟ # مناسب لطلاب الجامعات السورية الذين يريدون استخدام الإيميل الجامعي للنشاط الأكاديمي والمنصات البحثية. لا يضمن قبولك بأي برنامج طلابي مدفوع أو منحة؛ الأهلية النهائية دائماً حسب سياسة الجهة المقدمة. لا يغني عن الرجوع إلى المسار الإداري الرسمي داخل جامعتك. لماذا قسمنا الدليل إلى صفحتين؟ # لتسهيل الوصول السريع حسب نية البحث. لتقليل التشتت بين \u0026ldquo;الإجراء الإداري\u0026rdquo; و\u0026quot;الامتيازات/القيود\u0026quot;. لتحديث كل جزء بشكل مستقل عندما تتغير السياسات. CTA واضح للتحديثات # إذا بدك تحديثات فورية على أي منحة أو تغيير أهلية مرتبط بالإيميل الجامعي، تابع قناة التلغرام (رابط تتبع):\nTelegram Updates\nسجل التحديثات # 2026-04-13: إعادة هيكلة الدليل إلى مسارين، تحسين العناوين، إضافة مصادر رسمية وروابط داخلية. ","date":"13 April 2026","externalUrl":null,"permalink":"/tutorials-guides/university-email/","section":"Tutorials \u0026 Technical Guides","summary":"صفحة مرجعية تقسم الموضوع لمسارين واضحين: استخراج الإيميل الجامعي، ثم الفوائد والقيود مع مصادر رسمية.","title":"دليل الإيميل الجامعي في سوريا 2026: الاستخراج، الفوائد، والقيود","type":"tutorials-guides"},{"content":" فوائد الإيميل الجامعي في سوريا وحدوده: ما الذي يفتح فعلياً وما الذي يبقى مقيداً # آخر مراجعة للمحتوى والسياسات: 2026-04-13\nمهم: الأهلية تختلف بين مزود وآخر، وقد تتغير بدون إشعار.\nملخص سريع # الإيميل الجامعي يفيدك أكاديمياً في الهوية البحثية، والاعتماد المؤسسي، وبعض الفرص عند فتحها. الإيميل الجامعي وحده لا يضمن أي Student Plan مدفوع أو وصول تلقائي للمحتوى البحثي المغلق. ما الذي يفتحه الإيميل الجامعي غالباً؟ # الحالة ماذا يعني عملياً؟ ملاحظات حسابات باحث ربط الهوية الأكاديمية على ORCID وResearchGate وGoogle Scholar يفيد في الظهور المهني وتتبع الإنتاج البحثي تحقق مؤسسي إثبات الانتماء الجامعي في نماذج ومجتمعات أكاديمية ليس ضمان قبول نهائي لكل خدمة فرص دورية الاستفادة من عروض أو منح طلابية عند فتح بابها تعتمد على سياسة كل شركة وتاريخ الإطلاق ما الذي لا يضمنه الإيميل الجامعي؟ # الحالة لماذا قد تبقى مقيدة؟ التأثير العملي Student Plans مدفوعة متطلبات أهلية بلد/مؤسسة/اعتماد جامعي وسياسات امتثال قد يتم الرفض حتى مع إيميل جامعي الوصول للأوراق المدفوعة يتطلب اشتراكاً مؤسسياً (جامعة/مكتبة) واتفاقيات ترخيص لا يوجد وصول مجاني تلقائي بالفرد تفسير الفرق: نشر بحث vs عرض البحث # النشر العلمي يتم عبر المجلات وسياساتها التحريرية. منصات الاستعراض (مثل ResearchGate) تساعدك في عرض ملفك العلمي وبناء شبكة أكاديمية. لذلك: يمكن أن تنشر في مجلة دون أن يكون لديك حضور قوي على منصات الاستعراض، والعكس صحيح. مصادر رسمية للتحقق (External Citations) # أهلية الطالب والمنح التقنية # GitHub Education (Student Developer Pack):\nhttps://docs.github.com/en/education/about-github-education/github-education-for-students/apply-to-github-education-as-a-student GitHub Trade Controls (قيود الامتثال):\nhttps://docs.github.com/en/site-policy/other-site-policies/github-and-trade-controls Autodesk Education Overview:\nhttps://www.autodesk.com/education/edu-software/overview SOLIDWORKS Student Access:\nhttps://www.solidworks.com/product/students الوصول للأبحاث المدفوعة # Elsevier institutional access (ScienceDirect):\nhttps://www.elsevier.com/solutions/sciencedirect/for-libraries Wiley licensing and access for institutions:\nhttps://www.wiley.com/en-us/network/librarians FAQ مختصر # هل أقدر أنشر بحث من إيميل شخصي؟ # نعم، حسب متطلبات المجلة. الإيميل الجامعي يقوي الهوية الأكاديمية والارتباط المؤسسي لكنه ليس الشرط الوحيد للنشر.\nهل الإيميل الجامعي يكفي للحصول على Student Plan؟ # لا. هو عنصر داعم فقط، والقبول النهائي مرتبط بالأهلية والسياسات والامتثال لدى الجهة المقدمة.\nهل الإيميل الجامعي يعطي وصولاً تلقائياً لكل الأوراق المدفوعة؟ # لا. الوصول يعتمد على اشتراكات الجامعة واتفاقيات الترخيص مع الناشرين.\nCTA # إذا بدك تحديثات مستمرة على الفرص الجديدة وحالة الأهلية، تابع قناة التلغرام:\nTelegram Updates\nروابط داخلية مفيدة # كيفية استخراج الإيميل الجامعي خطوة بخطوة About Authenticity \u0026amp; Editorial Policy سجل التحديثات # 2026-04-13: تحويل الصفحة إلى صيغة \u0026ldquo;مسموح/مقيد\u0026rdquo; مع مصادر تحقق مباشرة. ","date":"13 April 2026","externalUrl":null,"permalink":"/tutorials-guides/university-email/university-email-benefits-limitations-syria/","section":"Tutorials \u0026 Technical Guides","summary":"دليل واضح بالمسموح وغير المتاح مع مصادر رسمية وروابط تحقق.","title":"فوائد الإيميل الجامعي في سوريا وحدوده: ما الذي يفتح فعلياً وما الذي يبقى مقيداً","type":"tutorials-guides"},{"content":"","date":"13 April 2026","externalUrl":null,"permalink":"/tags/%D9%81%D9%88%D8%A7%D8%A6%D8%AF-%D8%A7%D9%84%D8%A7%D9%8A%D9%85%D9%8A%D9%84-%D8%A7%D9%84%D8%AC%D8%A7%D9%85%D8%B9%D9%8A/","section":"Tags","summary":"","title":"فوائد الايميل الجامعي","type":"tags"},{"content":" كيفية استخراج الإيميل الجامعي في الجامعات السورية: دليل عملي خطوة بخطوة # آخر مراجعة: 2026-04-13\nالنتيجة المتوقعة: غالباً 20-30 دقيقة عند تجهيز الورقة مسبقاً.\nمهم: المسميات الإدارية قد تختلف بين الجامعات.\nما الذي ستحصل عليه من هذا الدليل؟ # المسار التنفيذي لاستخراج الإيميل الجامعي. قائمة تجهيز سريعة قبل الذهاب. أخطاء متكررة تؤخر المعاملة وكيف تتجنبها. قبل البدء: تحقق من نظام جامعتك # هذا الدليل مبني على تجربة تطبيقية مباشرة، لكن ترتيب المكاتب والتسميات الإدارية قد يختلف بين جامعة وأخرى. تأكد من المسار الرسمي داخل جامعتك أولاً.\nالوثيقة المطلوبة # كتاب رسمي موجه من الطالب إلى رئيس الجامعة لطلب تفعيل بريد إلكتروني جامعي.\nرابط القالب الجاهز (Google Docs):\nhttps://docs.google.com/document/d/1XEdyiosiO1y-egy4bS7IjotaAMiVq6xpBbkCNnQdfkk/edit?usp=drivesdk\nطريقة تعديل القالب # File -\u0026gt; Make a copy، أو File -\u0026gt; Download -\u0026gt; Microsoft Word. عدل الحقول التالية بدقة: رأس الصفحة: اسم الكلية بالعربية (يمين) وبالإنجليزية (يسار). سطر \u0026ldquo;عن طريق السيد عميد كلية \u0026hellip;\u0026rdquo;: اسم الكلية الصحيح. سطر \u0026ldquo;طالب في\u0026rdquo;: الاسم الثلاثي + القسم + الكلية. أسفل الصفحة: مقدم الطلب + التاريخ. خطوات استخراج الإيميل الجامعي # عمادة الكلية: توقيع وختم الكتاب. نائب رئيس الجامعة للدراسات العليا والبحث العلمي: ختم إضافي وإرجاع الكتاب لك. مركز الحاسب/الجهة التقنية المخولة: تسليم الكتاب الموقّع للحصول على: البريد الإلكتروني الجامعي كلمة مرور مؤقتة رابط الدخول الرسمي بعد أول تسجيل دخول، غيّر كلمة المرور فوراً.\nمخطط العملية (للتصفح السريع) # flowchart TD A[تحضير الكتاب] --\u003e B[توقيع وختم العمادة] B --\u003e C[ختم نائب رئيس الجامعة] C --\u003e D[تسليم لمركز الحاسب] D --\u003e E[استلام البريد + كلمة مرور مؤقتة] E --\u003e F[تسجيل أول دخول وتغيير كلمة المرور] Checklist سريع # قبل ما تروح # الكتاب الرسمي جاهز ومعدل. الاسم والكنية بالإنجليزية مكتوبين بصيغة موحدة. عندك نسخة رقمية ونسخة مطبوعة عند الحاجة. أثناء التقديم # تأكد من وجود التوقيع والختم المطلوبين. احتفظ بنسخة مصورة من الطلب قبل التسليم. بعد التفعيل # اختبر تسجيل الدخول فوراً. غيّر كلمة المرور. فعّل بريد استرجاع أو وسيلة أمان إن كانت متاحة. أخطاء شائعة تؤخر المعاملة # كتابة الاسم الإنجليزي بأكثر من صيغة. نسيان تعديل اسم الكلية في سطر التحويل عبر العمادة. تسليم طلب بلا ختم/توقيع مكتمل. تأجيل تغيير كلمة المرور بعد الاستلام. تنبيه مهني مهم قبل أي نشر علمي # لا تشارك بيانات بريدك الجامعي مع أي طرف. لا تقبل إضافة أسماء مؤلفين دون مساهمة بحثية فعلية. وثّق أي اتفاق إشرافي يخص Corresponding Author كتابةً. روابط مرتبطة داخل الموقع # فوائد الإيميل الجامعي وحدوده للسوريين قسم Tutorials \u0026amp; Guides Authenticity \u0026amp; Editorial Policy CTA # للتحديثات الجديدة حول المنح والأهلية، تابع قناة التلغرام:\nTelegram Updates\n","date":"13 April 2026","externalUrl":null,"permalink":"/tutorials-guides/university-email/how-to-get-university-email-syria/","section":"Tutorials \u0026 Technical Guides","summary":"دليل تنفيذي سريع لاستخراج الإيميل الجامعي في الجامعات السورية مع تنبيهات عملية مبنية على تجربة مباشرة.","title":"كيفية استخراج الإيميل الجامعي في الجامعات السورية: دليل عملي خطوة بخطوة","type":"tutorials-guides"},{"content":" Authenticity \u0026amp; Editorial Policy # Last updated: April 11, 2026\nThis policy explains how technical content on Mulham Fetna Blog is produced, edited, and published.\n1. Scope # This policy applies to technical content in:\nCourses Roadmaps Tutorials \u0026amp; Guides Other sections may follow different editorial styles unless stated on the page itself.\n2. Authenticity of technical work # The technical scenarios, decisions, troubleshooting paths, and implementation approaches published in the scoped sections are based on my own professional experience.\nThese materials are grounded in real projects, real constraints, and real debugging workflows, not synthetic examples created only for presentation.\n3. Use of LLM assistance # I use large language models (LLMs) as an editorial assistant to improve structure, readability, and clarity.\nLLM support may be used for:\nRewriting rough notes into clearer prose. Improving organization, headings, and flow. Language polishing and formatting consistency. LLMs are not used to replace my underlying technical judgment, authorship, or direct experience.\n4. Authorship and accountability # I remain responsible for all published technical claims, methods, and recommendations in the scoped sections.\nIf any inaccuracy is discovered, it will be corrected in subsequent revisions.\n5. Relationship to other policies # This page complements the Intellectual Property Policy, Privacy Policy, and Terms of Service.\n","date":"11 April 2026","externalUrl":null,"permalink":"/authenticity-editorial-policy/","section":"Authenticity \u0026 Editorial Policy","summary":"Authenticity \u0026 Editorial Policy # Last updated: April 11, 2026\nThis policy explains how technical content on Mulham Fetna Blog is produced, edited, and published.\n1. Scope # This policy applies to technical content in:\nCourses Roadmaps Tutorials \u0026 Guides Other sections may follow different editorial styles unless stated on the page itself.\n2. Authenticity of technical work # The technical scenarios, decisions, troubleshooting paths, and implementation approaches published in the scoped sections are based on my own professional experience.\nThese materials are grounded in real projects, real constraints, and real debugging workflows, not synthetic examples created only for presentation.\n3. Use of LLM assistance # I use large language models (LLMs) as an editorial assistant to improve structure, readability, and clarity.\nLLM support may be used for:\nRewriting rough notes into clearer prose. Improving organization, headings, and flow. Language polishing and formatting consistency. LLMs are not used to replace my underlying technical judgment, authorship, or direct experience.\n","title":"Authenticity \u0026 Editorial Policy","type":"authenticity-editorial-policy"},{"content":" جلسة استشارة تقنية # فكرة الخدمة # هذه الخدمة عبارة عن جلسة استشارية فردية بموضوع تختاره أنت، بشرط أن يكون ضمن نطاق خبراتي العملية الفعلية.\nهذه ليست جلسة عشوائية؛ هي جلسة مبنية على تشخيص دقيق لمشكلتك وسياقك.\nأمثلة على المحاور الممكنة # تخطيط مسار تعلم تقني واقعي (Data/AI/MLOps/Robotics). مراجعة مشروع تقني أو خطة تنفيذ. تشخيص مشكلة تقنية أو تنظيم خطوات حلها. قرارات انتقال مهني تقني مبنية على وضعك الحالي. ماذا يشمل الاجتماع؟ # فهم السياق الكامل قبل إعطاء أي توصية. تشخيص جذور المشكلة وليس فقط الأعراض. خطة تنفيذ واضحة بخطوات عملية بعد الجلسة. توصيات أدوات/مصادر مخصصة لحالتك. ما لا تشمل الخدمة # تنفيذ كامل للمشروع بدلًا عنك. استشارات خارج نطاق خبرتي الفعلية. وعود بنتائج غير واقعية. الحجز والدفع # السعر حسب الجهة الجغرافية: داخل سوريا: 25$ للجلسة. المغتربون/الخليج: 50$ إلى 75$ للجلسة. دولي: 100$ للجلسة. الدفع عبر ShamCash فقط. تأكيد الحجز يتطلب دفعًا كاملًا مسبقًا. أي حجز بدون تأكيد الدفع يتم إلغاؤه تلقائيًا. يسمح بإعادة جدولة مرة واحدة عند إشعار مسبق 24 ساعة. لا يوجد استرداد بعد تنفيذ الجلسة. تنبيه مهم: للمدفوعات الدولية، تواصل معي مباشرة على contact@mulhamfetna.com وسيتم التعامل يدويًا مؤقتًا إلى أن يصبح حجم الطلب مبررًا لنظام دفع آلي.\nخطوات الحجز (مهم) # حوّل الرسوم عبر ShamCash أولًا. احجز الموعد عبر Cal. أي حجز بدون تأكيد الدفع يتم إلغاؤه تلقائيًا. معلومات الدفع عبر ShamCash # Payment Identifier: d9baec7d91b344eb9c2d04363dd84664 للموبايل (مسح مباشر): رابط الحجز المباشر # Book your 1-hour meeting\nملاحظات قبل الحجز # لتكون الجلسة فعالة، أرسل قبل الموعد:\nوصفًا مختصرًا لمشكلتك/هدفك. روابط أو ملفات مساعدة (إن وجدت). النتيجة التي تريد الوصول إليها بعد الجلسة. الدعم وحل المشاكل # لأي مشكلة في الحجز أو الدفع، أرسل التفاصيل مع لقطات شاشة إلى: contact@mulhamfetna.com\nBooking Widget # English Snapshot # One-to-one technical consultation on your selected topic, limited to real expertise areas and focused on actionable outcomes.\nBooking \u0026amp; Payment # Pay first via ShamCash (Identifier: d9baec7d91b344eb9c2d04363dd84664). Direct booking link: https://cal.com/mulham-fetna/1-hour-meeting Pricing: $25 (Syria), $50–$75 (Diaspora/Gulf), $100 (International) per session. For international payments, contact: contact@mulhamfetna.com for manual handling. ","date":"11 April 2026","externalUrl":null,"permalink":"/services/technical-consultation-session/","section":"Mentorship \u0026 Consultations","summary":" جلسة استشارة تقنية # فكرة الخدمة # هذه الخدمة عبارة عن جلسة استشارية فردية بموضوع تختاره أنت، بشرط أن يكون ضمن نطاق خبراتي العملية الفعلية.\nهذه ليست جلسة عشوائية؛ هي جلسة مبنية على تشخيص دقيق لمشكلتك وسياقك.\nأمثلة على المحاور الممكنة # تخطيط مسار تعلم تقني واقعي (Data/AI/MLOps/Robotics). مراجعة مشروع تقني أو خطة تنفيذ. تشخيص مشكلة تقنية أو تنظيم خطوات حلها. قرارات انتقال مهني تقني مبنية على وضعك الحالي. ماذا يشمل الاجتماع؟ # فهم السياق الكامل قبل إعطاء أي توصية. تشخيص جذور المشكلة وليس فقط الأعراض. خطة تنفيذ واضحة بخطوات عملية بعد الجلسة. توصيات أدوات/مصادر مخصصة لحالتك. ما لا تشمل الخدمة # تنفيذ كامل للمشروع بدلًا عنك. استشارات خارج نطاق خبرتي الفعلية. وعود بنتائج غير واقعية. الحجز والدفع # السعر حسب الجهة الجغرافية: داخل سوريا: 25$ للجلسة. المغتربون/الخليج: 50$ إلى 75$ للجلسة. دولي: 100$ للجلسة. الدفع عبر ShamCash فقط. تأكيد الحجز يتطلب دفعًا كاملًا مسبقًا. أي حجز بدون تأكيد الدفع يتم إلغاؤه تلقائيًا. يسمح بإعادة جدولة مرة واحدة عند إشعار مسبق 24 ساعة. لا يوجد استرداد بعد تنفيذ الجلسة. تنبيه مهم: للمدفوعات الدولية، تواصل معي مباشرة على contact@mulhamfetna.com وسيتم التعامل يدويًا مؤقتًا إلى أن يصبح حجم الطلب مبررًا لنظام دفع آلي.\n","title":"Technical Consultation Session","type":"services"},{"content":" Intellectual Property Policy # Last updated: April 10, 2026\nThis policy explains ownership, permitted use, and licensing rules for content published on Mulham Fetna Blog.\n1. Ownership and scope # Unless otherwise stated, all original content is owned by Mulham Fetna and protected under applicable copyright and intellectual property laws.\nThis includes, without limitation:\nWritten articles, tutorials, guides, and page content. Images, graphics, logos, and visual brand assets. Course/service descriptions and educational materials. Downloadable resources and compiled media published on the site. Site structure and custom authored creative materials. 2. Permitted use (without separate approval) # You may:\nRead and view content for personal, non-commercial use. Share direct links to public pages on this website. Quote short excerpts with clear attribution and a source link. 3. Prohibited use (without prior written permission) # You may not:\nCopy, republish, mirror, distribute, or publicly display substantial content. Translate, adapt, remix, or create derivative works for publication or commercial use. Reuse images, graphics, branding, course material, or downloadable assets. Perform automated scraping, bulk extraction, or archival replication of site content. Use any site content for AI dataset creation, model training, fine-tuning, embedding generation, or synthetic content pipelines. Resell or redistribute any content in paid or ad-supported channels. 4. Code samples and licensing clarification # Where a specific code sample or file explicitly includes an open-source license notice (for example MIT), that license applies only to that specific code artifact.\nNo open license is granted by default to:\nWritten educational content and explanations. Visual/media assets (images, graphics, logos, screenshots). Course/program documentation and commercial training materials. 5. Commercial use and licensing requests # Commercial, institutional, or public reuse requires prior written permission.\nTo request permission, contact:\nEmail: contact@mulhamfetna.com Please include:\nExact content URL(s). Intended use case and publication channel. Commercial/non-commercial context. Timeframe and audience size. 6. Enforcement and infringement handling # Unauthorized use may result in one or more actions, including:\nWritten takedown demand. Platform/reporting requests (including copyright complaint workflows). Access restrictions where applicable. Further legal remedies as permitted by applicable law. If you believe your material appears on this site in error, send a notice to the same contact email with supporting details.\n7. Relationship to Terms and Privacy # This policy complements the Terms of Service and should be read together with the Privacy Policy.\n8. Policy updates # This policy may be updated at any time. The newest version is effective once published on this page with the revised “Last updated” date.\n","date":"10 April 2026","externalUrl":null,"permalink":"/intellectual-property/","section":"© Intellectual Property - All Rights Reserved","summary":"Intellectual Property Policy # Last updated: April 10, 2026\nThis policy explains ownership, permitted use, and licensing rules for content published on Mulham Fetna Blog.\n1. Ownership and scope # Unless otherwise stated, all original content is owned by Mulham Fetna and protected under applicable copyright and intellectual property laws.\nThis includes, without limitation:\nWritten articles, tutorials, guides, and page content. Images, graphics, logos, and visual brand assets. Course/service descriptions and educational materials. Downloadable resources and compiled media published on the site. Site structure and custom authored creative materials. 2. Permitted use (without separate approval) # You may:\nRead and view content for personal, non-commercial use. Share direct links to public pages on this website. Quote short excerpts with clear attribution and a source link. 3. Prohibited use (without prior written permission) # You may not:\nCopy, republish, mirror, distribute, or publicly display substantial content. Translate, adapt, remix, or create derivative works for publication or commercial use. Reuse images, graphics, branding, course material, or downloadable assets. Perform automated scraping, bulk extraction, or archival replication of site content. Use any site content for AI dataset creation, model training, fine-tuning, embedding generation, or synthetic content pipelines. Resell or redistribute any content in paid or ad-supported channels. 4. Code samples and licensing clarification # Where a specific code sample or file explicitly includes an open-source license notice (for example MIT), that license applies only to that specific code artifact.\n","title":"© Intellectual Property - All Rights Reserved","type":"intellectual-property"},{"content":" Privacy Policy # Last updated: April 10, 2026\nThis Privacy Policy explains how Mulham Fetna Blog collects, uses, and protects personal information when you visit this website or contact me for services.\n1. Who controls your data # Controller: Mulham Fetna Contact email: contact@mulhamfetna.com 2. Information collected # I may collect:\nContact information you send voluntarily (such as name, email address, and message content). Usage and device data through analytics tools, such as pages visited, approximate location, browser/device type, and referring source. Transactional information needed to provide mentorship services (for example session type, scheduling details, and service-related communication records). 3. Cookies and analytics # This website uses Google Analytics. Google Analytics may set cookies and process usage data to help measure traffic and improve site performance.\nYou can control cookies through your browser settings and may block or delete cookies at any time. Blocking cookies may affect some site functionality.\n4. Why data is used # Data is used to:\nRespond to inquiries and provide requested services. Manage mentorship communication and scheduling. Improve website content, structure, and user experience. Maintain site security, prevent abuse, and comply with legal obligations. 5. Legal basis (international/general approach) # Where applicable, processing is based on one or more of:\nConsent (for voluntary submissions). Contractual necessity (to deliver requested services). Legitimate interests (site operations, analytics, and security). Legal compliance (when required by law). 6. Data sharing # I do not sell personal data. Data may be shared only with trusted service providers needed to operate the site and services, such as:\nWebsite hosting/deployment providers. Analytics provider (Google Analytics). Communication and scheduling tools used to deliver services. These providers process data under their own policies and applicable law.\n7. Data retention # Personal data is retained only as long as needed for legitimate business, service, or legal purposes. When no longer needed, data is deleted or anonymized where feasible.\n8. International transfers # Because web infrastructure and providers may operate globally, your data may be processed outside your country. Reasonable safeguards are applied through reputable providers and standard contractual controls where available.\n9. Your rights # Depending on your location and applicable laws, you may request to:\nAccess personal data. Correct inaccurate data. Delete personal data. Restrict or object to processing. Receive a copy of your data (data portability). Withdraw consent where processing is based on consent. To submit a request, email: contact@mulhamfetna.com\n10. Children and minors # This site is not directed to children under 13. For users under 18 seeking services, parent/guardian consent may be required.\n11. Security # Reasonable technical and organizational measures are used to protect personal information. However, no internet transmission or storage system can be guaranteed 100% secure.\n12. Third-party links # This site may link to third-party websites. I am not responsible for their privacy practices. Review their privacy policies before submitting personal information.\n13. Changes to this policy # This Privacy Policy may be updated from time to time. The updated version will be posted on this page with a revised “Last updated” date.\n","date":"10 April 2026","externalUrl":null,"permalink":"/privacy/","section":"Privacy Policy","summary":"Privacy Policy # Last updated: April 10, 2026\nThis Privacy Policy explains how Mulham Fetna Blog collects, uses, and protects personal information when you visit this website or contact me for services.\n1. Who controls your data # Controller: Mulham Fetna Contact email: contact@mulhamfetna.com 2. Information collected # I may collect:\nContact information you send voluntarily (such as name, email address, and message content). Usage and device data through analytics tools, such as pages visited, approximate location, browser/device type, and referring source. Transactional information needed to provide mentorship services (for example session type, scheduling details, and service-related communication records). 3. Cookies and analytics # This website uses Google Analytics. Google Analytics may set cookies and process usage data to help measure traffic and improve site performance.\nYou can control cookies through your browser settings and may block or delete cookies at any time. Blocking cookies may affect some site functionality.\n4. Why data is used # Data is used to:\n","title":"Privacy Policy","type":"privacy"},{"content":" Terms of Service # Last updated: April 10, 2026\nThese Terms of Service (“Terms”) govern your use of Mulham Fetna Blog and related mentorship services. By using this website or booking services, you agree to these Terms.\n1. Owner and contact # Owner/Operator: Mulham Fetna Contact: contact@mulhamfetna.com 2. Scope of services # This website provides educational and professional content, and may offer paid mentorship services including academic and professional guidance.\nService details, formats, and prices may be updated at any time and become effective when published.\n3. Eligibility # You must be legally able to agree to these Terms. If you are under 18, parent or legal guardian consent is required before purchasing or receiving mentorship services. 4. Acceptable use # You agree not to:\nUse the website for unlawful, abusive, fraudulent, or harmful activity. Attempt unauthorized access to systems, infrastructure, or data. Copy, scrape, or redistribute protected content in violation of site policies. Misrepresent identity or submit false information. 5. Intellectual property # Unless otherwise stated, website content is owned by Mulham Fetna and protected by applicable intellectual property laws.\nUse of content is limited to personal, non-commercial viewing unless explicit written permission is granted.\n6. Payments, scheduling, and refunds # For paid mentorship services:\nPayment terms are communicated before service confirmation. No refunds are provided after a booked session has been delivered. Rescheduling is allowed with reasonable notice and subject to availability. Missed sessions without notice may be treated as completed sessions. 7. No guarantees # Educational and mentorship services are provided for guidance and skill development.\nNo guarantee is made regarding scholarships, admissions, job placement, or specific professional outcomes.\n8. Third-party services and links # The website may contain links or integrations from third-party providers (for example analytics, hosting, or communication tools).\nUse of third-party services is governed by their own terms and privacy policies.\n9. Disclaimer of warranties # The website and services are provided on an “as is” and “as available” basis to the fullest extent permitted by law.\nNo express or implied warranty is provided regarding uninterrupted availability, accuracy, or fitness for a specific purpose.\n10. Limitation of liability # To the fullest extent permitted by law, Mulham Fetna is not liable for indirect, incidental, special, consequential, or punitive damages arising from use of the website or services.\nWhere liability cannot be excluded by law, it is limited to the amount paid (if any) for the relevant service.\n11. Privacy # Use of personal information is governed by the Privacy Policy.\n12. Termination or suspension # Access to services may be limited, suspended, or terminated for violations of these Terms, misuse, fraud, abuse, or legal compliance reasons.\n13. Changes to these Terms # These Terms may be updated periodically. The updated version is effective when published on this page, with the revised “Last updated” date.\n14. Governing framework # These Terms are written for a general international audience. Local mandatory consumer and data protection laws may apply based on your country of residence.\n","date":"10 April 2026","externalUrl":null,"permalink":"/terms-of-service/","section":"Terms of Service","summary":"Terms of Service # Last updated: April 10, 2026\nThese Terms of Service (“Terms”) govern your use of Mulham Fetna Blog and related mentorship services. By using this website or booking services, you agree to these Terms.\n1. Owner and contact # Owner/Operator: Mulham Fetna Contact: contact@mulhamfetna.com 2. Scope of services # This website provides educational and professional content, and may offer paid mentorship services including academic and professional guidance.\nService details, formats, and prices may be updated at any time and become effective when published.\n3. Eligibility # You must be legally able to agree to these Terms. If you are under 18, parent or legal guardian consent is required before purchasing or receiving mentorship services. 4. Acceptable use # You agree not to:\nUse the website for unlawful, abusive, fraudulent, or harmful activity. Attempt unauthorized access to systems, infrastructure, or data. Copy, scrape, or redistribute protected content in violation of site policies. Misrepresent identity or submit false information. 5. Intellectual property # Unless otherwise stated, website content is owned by Mulham Fetna and protected by applicable intellectual property laws.\nUse of content is limited to personal, non-commercial viewing unless explicit written permission is granted.\n","title":"Terms of Service","type":"terms-of-service"},{"content":" Comprehensive Dual-Boot System Recovery \u0026amp; Maintenance Guide # Arch Linux \u0026amp; Windows 11 UEFI Systems # 1. System Architecture \u0026amp; Failure Modes # 1.1 UEFI Boot Mechanics # Modern dual-boot systems rely on UEFI NVRAM (Non-Volatile Random Access Memory) to store boot entries. Unlike Legacy BIOS (MBR), UEFI maintains a database of bootloaders independent of disk order. Common failure modes include:\nNVRAM Wipe: BIOS reset, SSD removal/reinstallation, or CMOS battery loss clears boot entries EFI Partition Misalignment: Moving/resizing partitions changes start sectors, breaking GRUB\u0026rsquo;s file references Filesystem Locks: Windows Fast Startup/hibernation marks NTFS as dirty, preventing Linux write access Kernel Absence: Incomplete mounts or partition moves resulting in missing /boot/vmlinuz-linux 1.2 Pre-Recovery Data Collection # Before manipulation, capture system state:\n# Boot from Arch Live ISO sudo fdisk -l /dev/sda # Partition table geometry sudo lsblk -f # Filesystem types and UUIDs sudo blkid # Detailed partition attributes efibootmgr -v # Current UEFI boot entries (if available) 2. Bootloader Recovery \u0026amp; GRUB Restoration # 2.1 Scenario A: Lost Boot Entries (Post-BIOS Reset) # Symptoms: System boots directly to Windows; no GRUB menu; missing Linux entry in firmware boot menu.\nRoot Cause: UEFI NVRAM entries deleted while EFI partitions remain intact.\nRecovery Procedure:\nBoot Environment Preparation\n# Identify partitions (adjust device identifiers as needed) sudo mount /dev/sda7 /mnt # Arch root partition sudo mount /dev/sda6 /mnt/boot/efi # Arch EFI partition sudo mount /dev/sda8 /mnt/home # Optional: home partition # Virtual filesystem binding for chroot sudo mount --bind /dev /mnt/dev sudo mount --bind /proc /mnt/proc sudo mount --bind /sys /mnt/sys Chroot Entry\nsudo arch-chroot /mnt # Alternatively, manual method: # chroot /mnt /bin/bash GRUB Reinstallation\n# Re-register GRUB with UEFI firmware grub-install --target=x86_64-efi \\ --efi-directory=/boot/efi \\ --bootloader-id=GRUB \\ --recheck # Regenerate configuration with OS-prober for Windows detection grub-mkconfig -o /boot/grub/grub.cfg Boot Priority Restoration\n# Verify entry creation efibootmgr -v # Set GRUB as first boot option (replace XXXX with boot number) efibootmgr -o XXXX,YYYY,ZZZZ 2.2 Scenario B: GRUB Rescue Prompt (Post-Partition Move) # Symptoms: grub rescue\u0026gt; prompt; \u0026ldquo;unknown filesystem\u0026rdquo; errors; partition resized/moved recently.\nRoot Cause: GRUB\u0026rsquo;s core.img references incorrect block addresses after partition geometry changes.\nExtended Recovery:\nManual Boot (Temporary)\n# Identify partition containing /boot ls (hd0,gpt6)/ # Set root and prefix set root=(hd0,gpt6) set prefix=(hd0,gpt6)/boot/grub insmod normal normal Complete Restoration via Live ISO\n# Mount partitions sudo mount /dev/sda6 /mnt # Root sudo mount /dev/sda5 /mnt/boot/efi # EFI (new location after move) sudo arch-chroot /mnt # Critical: Update fstab for new UUIDs if partitions were reformatted nano /etc/fstab # Reinstall GRUB to new EFI location grub-install --target=x86_64-efi --efi-directory=/boot/efi --bootloader-id=GRUB # Kernel restoration (essential if /boot was unmounted during moves) pacman -S linux linux-lts # Reinstall kernel packages mkinitcpio -P # Regenerate initramfs for all kernels # Final configuration grub-mkconfig -o /boot/grub/grub.cfg UEFI Entry Cleanup\n# Remove stale entries (optional) efibootmgr -b XXXX -B # Delete old entry by number # Create new entry manually if automated creation fails efibootmgr -c -d /dev/sda -p 5 -l /EFI/GRUB/grubx64.efi -L \u0026#34;GRUB\u0026#34; 2.3 BIOS Configuration Requirements # Access firmware settings (F2/Del during POST) and verify:\nBoot Mode: UEFI only (disable Legacy/CSM) Secure Boot: Disabled (unless using signed kernels/shim) SATA Mode: AHCI (not RAID/Intel RST) Fast Boot: Disabled for troubleshooting (prevents USB initialization) 3. Permanent Filesystem Configuration # 3.1 NTFS Data Partition Integration # Objective: Mount Windows data partitions at boot with full read-write access for regular users.\nImplementation:\nPrerequisites\nsudo pacman -S ntfs-3g fuse2 # FUSE-based NTFS driver UUID Identification\nsudo blkid /dev/sda3 # Output: UUID=\u0026#34;01DC1F4F8BC938D0\u0026#34; TYPE=\u0026#34;ntfs\u0026#34; LABEL=\u0026#34;DATA\u0026#34; Mount Point Creation\nsudo mkdir -p /mnt/data sudo chown $USER:$USER /mnt/data # Pre-set ownership fstab Configuration\nsudo cp /etc/fstab /etc/fstab.bak # Backup # Edit configuration sudo nano /etc/fstab # Add entry: UUID=01DC1F4F8BC938D0 /mnt/data ntfs-3g defaults,uid=1000,gid=1000,umask=000,noatime,nofail 0 0 Parameter Analysis:\nuid=1000,gid=1000: Maps all files to your user/group ID (verify with id command) umask=000: Full permissions (rwX) for owner; adjust to umask=022 for read-only others noatime: Prevents write amplification on SSDs (NTFS journal overhead) nofail: Prevents boot failure if drive is missing/disconnected Activation\nsudo systemctl daemon-reload sudo mount -a findmnt --verify # Validate syntax 3.2 Alternative: NTFS3 Kernel Driver (Modern Approach) # For kernel 5.15+, the native ntfs3 driver offers improved performance:\n# In /etc/fstab: UUID=XXXX /mnt/data ntfs3 uid=1000,gid=1000,umask=000,nofail 0 0 Note: ntfs3 lacks NTFS compression support; use ntfs-3g if compressing files from Linux.\n4. Cross-Platform Filesystem Troubleshooting # 4.1 NTFS Lock States (Windows Hibernation) # Symptoms: Mount error \u0026ldquo;The disk contains an unclean file system\u0026rdquo;; read-only mount despite fstab settings; Operation not permitted on write.\nMechanism: Windows Fast Startup creates hiberfil.sys, marking the NTFS journal as in-use.\nResolution Hierarchy:\nMethod 1: Windows-side Permanent Fix (Recommended) # Disable Fast Startup in Windows:\nControl Panel → Power Options → Choose what power buttons do → Change settings unavailable → Uncheck Turn on fast startup Method 2: Linux Force Mount (Data Risk) # # Removes hibernation file (destroys unsaved Windows session) sudo mount -t ntfs-3g -o remove_hiberfile /dev/sda3 /mnt/data Method 3: Repair Utilities # # Clear dirty bit (limited effectiveness) sudo ntfsfix /dev/sda3 # Full Windows repair (requires Windows PE/Recovery) # Boot Windows Recovery → Command Prompt chkdsk D: /f /r /x # /f: Fix errors | /r: Recover bad sectors | /x: Force dismount 4.2 Busy Mount Resolution # When unmount fails with \u0026ldquo;target is busy\u0026rdquo;:\n# Identify blocking processes sudo fuser -m /mnt/data sudo lsof | grep /mnt/data # Graceful termination sudo kill -15 \u0026lt;PID\u0026gt; # Force unmount (lazy detach) sudo umount -l /mnt/data 4.3 Partition Recovery (Advanced) # If partition table corruption occurs during resizing:\n# Install recovery tools sudo pacman -S testdisk # Interactive recovery sudo testdisk /dev/sda # Select: Intel/PC → Analyze → Quick Search → Deeper Search → Write 5. Diagnostic Reference Matrix # Symptom Diagnostic Command Root Cause Solution GRUB rescue prompt ls in GRUB shell Moved/deleted EFI partition Section 2.2 Boots directly to Windows efibootmgr -v NVRAM entry lost Section 2.1 mount: unknown filesystem type 'ntfs' lsblk -f Missing ntfs-3g Section 3.1 NTFS read-only dmesg | grep ntfs Windows hibernation lock Section 4.1 Missing kernel in GRUB ls /boot in chroot Unmounted /boot during install Section 2.2 (kernel reinstall) failed to mount /boot/efi blkid vs /etc/fstab UUID mismatch after format Update fstab UUIDs 6. Validation Checklist # Post-recovery verification steps:\nefibootmgr -v shows GRUB entry with correct path \\EFI\\GRUB\\grubx64.efi sudo grub-mkconfig detects both Linux and Windows bootloaders Reboot test: System presents GRUB menu (timeout ≥ 5 seconds) Windows boot successful, then Linux boot successful (bi-directional) findmnt /mnt/data shows NTFS partition with rw flags Test write: touch /mnt/data/linux_test_file succeeds without sudo Windows restart does not re-lock NTFS (Fast Startup disabled) 7. Prevention \u0026amp; Maintenance # UEFI Entry Backup\nsudo efibootmgr -v \u0026gt; /boot/efi/efi_backup_$(date +%Y%m%d).txt fstab Immutability After confirmed working configuration:\nsudo chattr +i /etc/fstab # Prevent accidental modification sudo chattr -i /etc/fstab # When changes needed Pre-Resize Protocol Before partition manipulation:\nDisable Windows Fast Startup Disable BIOS Secure Boot Create GRUB rescue USB with grub-install --removable Monitoring\n# Weekly filesystem check sudo ntfsfix -n /dev/sda3 # No-action check for dirty bit This guide consolidates bootloader recovery, permanent mounting configuration, and cross-platform filesystem troubleshooting into a single reference for maintaining complex dual-boot environments. All procedures assume UEFI/GPT partitioning; adapt device identifiers (sda, nvme0n1, etc.) to your specific hardware topology.\n8. Frequently Asked Questions (FAQ) # 8.1 UEFI Boot Architecture # Q1: Why does removing the SSD physically clear UEFI boot entries when the data is still intact? A: UEFI stores boot entries in NVRAM (non-volatile memory) on the motherboard, not on the disk. These entries contain pointers to EFI files (like \\EFI\\GRUB\\grubx64.efi). When you remove the SSD, the firmware detects the missing device and often purges invalid entries during POST. Additionally, resetting BIOS/CMOS clears NVRAM entirely. The EFI partition itself remains unharmed, but the firmware \u0026ldquo;forgets\u0026rdquo; the path to the bootloader.\nQ2: What is the fundamental difference between Legacy BIOS (MBR) and UEFI boot modes? A: Legacy BIOS uses the Master Boot Record (first 512 bytes of disk) containing stage-1 boot code that chainloads stage-2 from disk sectors. UEFI uses a FAT32-formatted EFI System Partition (ESP) containing .efi executable files. UEFI maintains a boot manager database in NVRAM, supports GPT partitioning (no 2TB limit), and offers Secure Boot capabilities. Legacy modifies disk sectors directly; UEFI launches applications from the filesystem.\nQ3: Can I use a single EFI partition for both Windows and Linux? A: Yes, but it is not recommended for complex setups. Both Windows and Linux can coexist in /boot/efi/EFI/Microsoft and /boot/efi/EFI/GRUB respectively. However, Windows Update occasionally \u0026ldquo;cleans\u0026rdquo; the EFI partition and may remove non-Microsoft bootloaders. Separate EFI partitions (Windows on sda1, Linux on sda6) provide isolation but require manual boot entry management via efibootmgr.\nQ4: Why does GRUB rescue appear immediately after I resized/moved partitions with GParted? A: GRUB\u0026rsquo;s core.img contains absolute block addresses (not filesystem paths) for stage-2 files. When you move a partition, the start sector changes, but GRUB\u0026rsquo;s embedded pointer still references the old physical location. The rescue prompt appears because GRUB cannot locate its normal.mod or configuration files at the expected disk offsets.\nQ5: How can I manually boot my system from the GRUB rescue prompt without a live USB? A: You can manually chainload if you know the partition layout:\ngrub rescue\u0026gt; ls (hd0,gpt6)/ # List root contents grub rescue\u0026gt; set root=(hd0,gpt6) grub rescue\u0026gt; set prefix=(hd0,gpt6)/boot/grub grub rescue\u0026gt; insmod normal grub rescue\u0026gt; normal grub rescue\u0026gt; insmod linux grub rescue\u0026gt; linux /boot/vmlinuz-linux root=/dev/sda6 grub rescue\u0026gt; initrd /boot/initramfs-linux.img grub rescue\u0026gt; boot This temporarily boots the system so you can permanently reinstall GRUB.\n8.2 Chroot \u0026amp; System Recovery # Q6: What is the difference between chroot and arch-chroot, and when must I use the manual method? A: arch-chroot is a wrapper that automatically mounts /dev, /proc, /sys, /run, and binds /etc/resolv.conf before entering the chroot. Use manual chroot when arch-chroot fails or on non-Arch systems:\nmount --types proc /proc /mnt/proc mount --rbind /sys /mnt/sys mount --make-rslave /mnt/sys mount --rbind /dev /mnt/dev mount --make-rslave /mnt/dev mount --bind /run /mnt/run mount --make-slave /mnt/run cp /etc/resolv.conf /mnt/etc/ chroot /mnt /bin/bash Q7: Why does grub-install fail with \u0026ldquo;EFI variables cannot be set on this system\u0026rdquo;? A: This occurs when booted in Legacy BIOS mode or when the EFI filesystem drivers are not loaded. Inside chroot, you must bind-mount the efivarfs:\nmount -t efivarfs efivarfs /sys/firmware/efi/efivars Also ensure you\u0026rsquo;re booted the USB in UEFI mode (not \u0026ldquo;Legacy USB Support\u0026rdquo;).\nQ8: What happens if my /boot directory was unmounted during a kernel upgrade, and how do I detect this? A: The kernel image (vmlinuz-linux) and initramfs are written to the root filesystem\u0026rsquo;s /boot directory instead of the EFI partition. GRUB configuration points to the EFI partition, so it cannot find the kernel. Symptoms: GRUB menu shows only Windows. Detection:\n# In chroot ls -la /boot/efi/EFI/GRUB/grubx64.efi # Should exist ls -la /boot/vmlinuz-linux # If missing here but pacman says installed, /boot wasn\u0026#39;t mounted Fix: Mount /boot correctly, reinstall kernel packages pacman -S linux, regenerate GRUB config.\nQ9: How do I recover if I accidentally formatted my Linux EFI partition? A: Boot live ISO, mount root and home (but not the missing EFI). Create new FAT32 partition, mount it at /mnt/boot/efi, then:\narch-chroot /mnt pacman -S grub efibootmgr grub-install --target=x86_64-efi --efi-directory=/boot/efi --bootloader-id=GRUB grub-mkconfig -o /boot/grub/grub.cfg You will need to manually copy grubx64.efi to the new partition; GRUB reinstall handles this.\nQ10: Why does mkinitcpio -P fail with \u0026ldquo;command not found\u0026rdquo; inside chroot? A: The chroot environment lacks the PATH variable or base-devel tools. Export PATH explicitly:\nexport PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin Or ensure base and mkinitcpio packages are installed: pacman -S base mkinitcpio linux.\n8.3 Filesystem \u0026amp; fstab Configuration # Q11: How do I determine my actual user UID and GID for fstab configuration? A: Run id in terminal:\nuid=1000(username) gid=1000(username) groups=1000(username),3(sys),90(network),98(power) The first number (1000) is your UID. For system consistency, use 1000 for the primary user. If you have multiple users, create a shared group or use gid=users (typically 100).\nQ12: Why is my NTFS partition mounted read-only despite using ntfs-3g and correct fstab options? A: Three common causes:\nWindows hibernation: See Section 4.1. Check with ntfs-3g.secaudit /dev/sda3 | grep hibernated Filesystem errors: NTFS marked dirty. Run sudo ntfsfix /dev/sda3 or Windows chkdsk Mount precedence: An earlier mount (systemd, udisks) mounted it read-only before fstab applied. Check findmnt /mnt/data and unmount before mount -a Q13: Should I use ntfs-3g (FUSE) or the newer ntfs3 (kernel driver)? A: Use ntfs-3g if you need:\nFull POSIX permission support (chmod/chown) Windows compression (NTFS-LZNT1) support Proven stability for critical writes Use ntfs3 (kernel 5.15+) if you want:\nBetter performance (native kernel vs FUSE overhead) Lower CPU usage Simple mount options (no uid=/gid= needed if using ACLs) Q14: What are the risks of using remove_hiberfile mount option on a production Windows system? A: This deletes hiberfil.sys, destroying:\nUnsaved work in hibernated Windows sessions Windows Fast Startup cache (causing slower subsequent boots) Hybrid sleep data Always attempt Windows full shutdown (Shift+Shutdown) or powercfg /hibernate off from Windows first. Q15: Why does mount -a report \u0026ldquo;can\u0026rsquo;t find UUID=XXXX\u0026rdquo; when blkid shows the UUID exists? A: This indicates the kernel hasn\u0026rsquo;t updated its partition table cache after a disk modification. Run:\npartprobe /dev/sda # or blockdev --rereadpt /dev/sda If the device is busy, you may need to reboot or use partprobe -s.\n8.4 Cross-Platform Issues # Q16: Can I disable Windows Fast Startup from within Linux without booting Windows? A: No. Fast Startup is a Windows registry and powercfg setting. However, you can force a \u0026ldquo;full shutdown\u0026rdquo; from Windows command line (if you can access Windows RE):\nshutdown /s /f /t 0 The only Linux-side workaround is remove_hiberfile (destructive) or mounting read-only.\nQ17: Why does Windows CHKDSK run automatically every time I boot Windows after using Linux? A: Windows detects the \u0026ldquo;dirty bit\u0026rdquo; set by ntfsfix or forced mounts. While ntfsfix clears some flags, it may set others that Windows interprets as \u0026ldquo;volume modified by external OS.\u0026rdquo; Allow CHKDSK to run once; it will reset internal consistency markers.\nQ18: Is it safe to resize NTFS partitions from Linux using GParted? A: GParted uses ntfsresize (part of ntfs-3g suite) which is generally safe, but:\nAlways defragment Windows first (defrag C: /O) Disable Windows pagefile and hibernation during resize Never resize a \u0026ldquo;dirty\u0026rdquo; NTFS partition (run chkdsk first) Have Windows repair media ready (rarely, Windows may need to reassign drive letters) Q19: How do I access my Linux ext4 partition from Windows for data recovery? A: Use Windows Subsystem for Linux (WSL2) with wsl --mount --type ext4, or install third-party drivers like Ext2Fsd (discontinued, risky) or Paragon ExtFS. For recovery, boot Linux live ISO instead—native Linux tools are safer than Windows drivers for ext4.\nQ20: Can I use dd to clone my dual-boot disk to a new SSD? A: Yes, but post-clone you must:\n# On the new disk sgdisk -G /dev/sdb # Randomize GUID to avoid conflicts efibootmgr -c -d /dev/sdb -p 1 -L \u0026#34;GRUB Clone\u0026#34; -l \u0026#39;\\EFI\\GRUB\\grubx64.efi\u0026#39; Clone with dd if=/dev/sda of=/dev/sdb bs=4M status=progress or use ddrescue for failing disks.\n8.5 Advanced Troubleshooting # Q21: What is Secure Boot, why does it prevent GRUB installation, and can I use it with Arch? A: Secure Boot verifies cryptographic signatures on EFI binaries against Microsoft\u0026rsquo;s UEFI CA. Unsigned GRUB (default Arch) is rejected. Solutions:\nDisable Secure Boot (easiest) Use shim and preloader to chainload unsigned GRUB (complex) Sign your own kernels and GRUB with MOK (Machine Owner Key) using sbsigntools Q22: How do I create a Windows 10/11 bootable USB from Arch Linux for repairing BCD/CHKDSK? A: Use woeusb or manual formatting:\n# Replace /dev/sdX with your USB (not partition) sudo wipefs -a /dev/sdX sudo fdisk /dev/sdX # Create GPT, partition 1 type Microsoft basic data sudo mkfs.fat -F32 /dev/sdX1 sudo mkdir -p /mnt/winusb sudo mount /dev/sdX1 /mnt/winusb # Mount Windows ISO sudo mount -o loop win11.iso /mnt/iso sudo cp -r /mnt/iso/* /mnt/winusb/ sudo sync Alternatively, use Ventoy for multi-ISO USB.\nQ23: Why does my system boot to a black screen after selecting Arch in GRUB? A: Common causes:\nMissing GPU drivers: If you installed NVIDIA/AMD drivers but boot generic kernel without initramfs updates Incorrect root UUID: Kernel parameter root= points to wrong partition (check /etc/default/grub and grub.cfg) Fast Boot enabled in BIOS: GPU not initialized properly; disable in UEFI settings Q24: How do I restore Windows Boot Manager as the default bootloader temporarily? A: From Windows:\nbcdedit /set {bootmgr} path \\EFI\\Microsoft\\Boot\\bootmgfw.efi From Linux:\nefibootmgr -n 000X # Set Windows entry as next boot only # or permanently: efibootmgr -o 000X,000Y # Windows first, GRUB second Q25: What is the difference between UUID and PARTUUID, and which should I use in fstab? A:\nUUID: Filesystem identifier (e.g., 01DC1F4F8BC938D0). Changes when you reformat the partition. Use for data partitions. PARTUUID: GPT partition table entry identifier (e.g., 12345678-1234-1234-1234-123456789abc). Persists across formats. Use for root/boot in /etc/fstab to avoid boot failures after accidental reformats. Syntax: PARTUUID=xxxx-yyyy vs UUID=xxxx. Q26: How do I backup and restore my GPT partition table (not the data)? A: Backup:\nsgdisk -b gpt_backup.bin /dev/sda Restore (destructive):\nsgdisk -l gpt_backup.bin /dev/sda partprobe /dev/sda Store gpt_backup.bin on external media. This preserves partition layout but not filesystem contents.\nQ27: Why does findmnt --verify warn about \u0026ldquo;ntfs-3g: unknown filesystem type\u0026rdquo;? A: The verification tool checks kernel filesystem support, but ntfs-3g is a FUSE module, not a kernel driver. This warning is cosmetic; the mount will succeed if ntfs-3g is installed. For kernel-native ntfs3, the warning disappears.\nQ28: Can I convert my existing Legacy BIOS install to UEFI without reinstalling? A: Yes, if your motherboard supports UEFI:\nBoot live ISO in UEFI mode Create FAT32 EFI partition (~512MB) Mount root and EFI arch-chroot /mnt pacman -S grub efibootmgr grub-install --target=x86_64-efi --efi-directory=/boot/efi --bootloader-id=GRUB Update fstab with new EFI UUID Reboot and switch BIOS to UEFI mode Q29: What should I do if arch-chroot fails with \u0026ldquo;failed to run command \u0026lsquo;/bin/bash\u0026rsquo;: No such file or directory\u0026rdquo;? A: This indicates the root partition was mounted without the actual system files (possibly mounted the EFI partition as root by mistake), or the bash binary is corrupted. Verify:\nls /mnt/bin/bash # If missing, you mounted wrong partition. If exists but error persists: mount -t proc proc /mnt/proc mount -t sysfs sys /mnt/sys mount -o bind /dev /mnt/dev # Then try chroot with explicit path: chroot /mnt /usr/bin/bash Q30: How do I handle a system where both OSes show in GRUB, but Windows crashes with \u0026ldquo;INACCESSIBLE_BOOT_DEVICE\u0026rdquo; after I fixed Linux? A: This occurs when GRUB\u0026rsquo;s chainloader passes incorrect firmware parameters to Windows, or when the Windows EFI partition was accidentally modified. Fix from Windows Recovery:\nbootrec /fixboot bootrec /scanos bootrec /rebuildbcd Or from Linux, ensure os-prober correctly identifies Windows EFI partition:\npacman -S os-prober grub-mkconfig -o /boot/grub/grub.cfg 9. Official System Documentation \u0026amp; References # 9.1 Arch Linux Ecosystem # Arch Wiki (Primary Authority)\nDual boot with Windows: https://wiki.archlinux.org/title/Dual_boot_with_Windows GRUB: https://wiki.archlinux.org/title/GRUB Fstab: https://wiki.archlinux.org/title/Fstab NTFS-3G: https://wiki.archlinux.org/title/NTFS-3G NTFS (Kernel Driver): https://wiki.archlinux.org/title/NTFS Chroot: https://wiki.archlinux.org/title/Chroot Unified Extensible Firmware Interface: https://wiki.archlinux.org/title/Unified_Extensible_Firmware_Interface Partitioning: https://wiki.archlinux.org/title/Partitioning System Recovery: https://wiki.archlinux.org/title/Installation_guide#Chroot Arch Manual Pages\nman 5 fstab – Static filesystem information man 8 mount – Mount filesystems man 8 ntfs-3g – NTFS FUSE driver manual man 8 grub-install – Install GRUB on a device man 8 efibootmgr – Manage UEFI Boot Entries man 8 mkinitcpio – Create initial ramdisk environments 9.2 GNU \u0026amp; Linux Kernel Documentation # GNU GRUB Manual\nUEFI Installation: https://www.gnu.org/software/grub/manual/grub/html_node/Installing-GRUB-using-grub_002dinstall.html GRUB Rescue: https://www.gnu.org/software/grub/manual/grub/html_node/GRUB-only-offers-a-rescue-shell.html Linux Kernel Documentation\nNTFS3 Driver (Kernel 5.15+): https://www.kernel.org/doc/html/latest/filesystems/ntfs3.html EFI Variables: https://www.kernel.org/doc/html/latest/admin-guide/efi.html Filesystem Hierarchy Standard: https://refspecs.linuxfoundation.org/FHS_3.0/fhs/index.html 9.3 Microsoft Technical Documentation # Windows Hardware Developer Center\nUEFI Firmware: https://docs.microsoft.com/en-us/windows-hardware/drivers/bringup/uefi-firmware Boot Configuration Data (BCD): https://docs.microsoft.com/en-us/windows-hardware/manufacture/desktop/bcd-boot-options-reference CHKDSK Command: https://docs.microsoft.com/en-us/windows-server/administration/windows-commands/chkdsk Fast Startup Technical Details: https://docs.microsoft.com/en-us/windows-hardware/design/device-experiences/powercfg-command-line-options Windows Recovery Environment (WinRE)\nBootrec.exe Tool: https://docs.microsoft.com/en-us/windows-hardware/manufacture/desktop/bootsect-command-line-options BCDEdit: https://docs.microsoft.com/en-us/windows-server/administration/windows-commands/bcdedit 9.4 UEFI Specifications # UEFI Forum\nUEFI Specification 2.9B (PDF): https://uefi.org/specifications Secure Boot: https://uefi.org/revocationlistfile GPT Partitioning: Chapter 5 of UEFI Specification (GUID Partition Table) 9.5 Filesystem \u0026amp; Recovery Tools # NTFS-3G Project\nDocumentation: https://github.com/tuxera/ntfs-3g/wiki Advanced Options: https://www.tuxera.com/community/open-source-ntfs-3g/ GParted\nResizing NTFS: https://gparted.org/features.php Documentation: https://gparted.org/display-doc.php TestDisk \u0026amp; PhotoRec\nNTFS Repair Guide: https://www.cgsecurity.org/wiki/NTFS_Boot_sector_recovery Partition Table Recovery: https://www.cgsecurity.org/wiki/TestDisk 9.6 Dell Precision-Specific Resources # Dell Support\nPrecision 3551 UEFI Configuration: https://www.dell.com/support/manuals/precision-3551-workstation BIOS Setup Guide: https://www.dell.com/support/kbdoc/000124211 10. Version History \u0026amp; Change Log # Version Date Changes 1.0 2026-03-31 Initial comprehensive guide combining bootloader recovery, NTFS mounting, and UEFI troubleshooting Disclaimer: This guide involves low-level disk and firmware operations. Always backup critical data before executing partition table modifications, bootloader reinstallations, or filesystem repairs. The procedures assume x86_64 architecture with GPT partitioning; adaptations may be required for ARM64 or MBR systems.\n","date":"31 March 2026","externalUrl":null,"permalink":"/tutorials-guides/dual-boot-system-recovery-maintenance-guide/","section":"Tutorials \u0026 Technical Guides","summary":"Comprehensive Dual-Boot System Recovery \u0026 Maintenance Guide # Arch Linux \u0026 Windows 11 UEFI Systems # 1. System Architecture \u0026 Failure Modes # 1.1 UEFI Boot Mechanics # Modern dual-boot systems rely on UEFI NVRAM (Non-Volatile Random Access Memory) to store boot entries. Unlike Legacy BIOS (MBR), UEFI maintains a database of bootloaders independent of disk order. Common failure modes include:\nNVRAM Wipe: BIOS reset, SSD removal/reinstallation, or CMOS battery loss clears boot entries EFI Partition Misalignment: Moving/resizing partitions changes start sectors, breaking GRUB’s file references Filesystem Locks: Windows Fast Startup/hibernation marks NTFS as dirty, preventing Linux write access Kernel Absence: Incomplete mounts or partition moves resulting in missing /boot/vmlinuz-linux 1.2 Pre-Recovery Data Collection # Before manipulation, capture system state:\n# Boot from Arch Live ISO sudo fdisk -l /dev/sda # Partition table geometry sudo lsblk -f # Filesystem types and UUIDs sudo blkid # Detailed partition attributes efibootmgr -v # Current UEFI boot entries (if available) 2. Bootloader Recovery \u0026 GRUB Restoration # 2.1 Scenario A: Lost Boot Entries (Post-BIOS Reset) # Symptoms: System boots directly to Windows; no GRUB menu; missing Linux entry in firmware boot menu.\n","title":"Dual-Boot System Recovery \u0026 Maintenance Guide","type":"tutorials-guides"},{"content":" Comprehensive ESP32 Arduino Core Installation Guide # Overview: Two Installation Methods # Method Best For Pros Cons Boards Manager Beginners, stability One-click install, version management, automatic updates Slight delay for newest features Manual/Git Clone Developers, beta testers Latest commits, instant updates, offline capable Manual updates, potential instability Method 1: Boards Manager (Recommended for Most Users) # Universal Prerequisites # Arduino IDE 1.8.19+ or Arduino IDE 2.3+ (IDE 2.x recommended) Python 3.7+ (usually bundled with IDE 2.x) USB Cable: Must be data-capable, not charge-only Step 1: Add ESP32 Board Manager URL # Arduino IDE 2.x:\nFile → Preferences (or Ctrl+,) In \u0026ldquo;Additional boards manager URLs\u0026rdquo; add: https://raw.githubusercontent.com/espressif/arduino-esp32/gh-pages/package_esp32_index.json Click OK Arduino IDE 1.8.x:\nFile → Preferences Same URL as above If multiple URLs, separate with commas Step 2: Install the Core # Tools → Board → Boards Manager\u0026hellip; Search for \u0026ldquo;ESP32\u0026rdquo; by Espressif Systems Select version (latest stable recommended, e.g., 3.0.x or 2.0.x) Click Install (downloads ~500MB-1GB of toolchains) Step 3: Select Your Board # Tools → Board → ESP32 Arduino → Select your specific board:\nESP32 Dev Module (generic, works for most ESP32-WROOM boards) ESP32-S2/S3/C3 Dev Module (for specific variants) ESP32-CAM, TTGO T-Display, etc. (for specific hardware) Method 2: Manual Installation (Git Clone) # Use this for development versions, unreleased fixes, or offline environments.\nWindows (PowerShell/Command Prompt) # # 1. Create directory structure mkdir %USERPROFILE%\\Documents\\Arduino\\hardware\\espressif cd %USERPROFILE%\\Documents\\Arduino\\hardware\\espressif # 2. Clone repository (shallow clone for faster download) git clone --recursive https://github.com/espressif/arduino-esp32.git esp32 cd esp32 # 3. Optional: Checkout specific version git checkout 3.0.7 # or 2.0.17 for legacy # 4. Download tools (requires Python) cd tools python get.py # 5. Install Python dependencies (if get.py fails) pip install pyserial Windows-Specific Notes:\nIf python command not found, use py or python3 Path might be C:\\Users\\\u0026lt;Username\u0026gt;\\Arduino\\... if Documents folder is redirected Windows Defender may flag toolchain downloads as false positives—allow them macOS (Terminal) # # 1. Install prerequisites (if not present) brew install python git # or use macOS system python3 # 2. Create directories mkdir -p ~/Documents/Arduino/hardware/espressif cd ~/Documents/Arduino/hardware/espressif # 3. Clone with submodules git clone --recursive https://github.com/espressif/arduino-esp32.git esp32 cd esp32 # 4. Checkout stable (optional) git checkout 3.0.7 # 5. Get tools cd tools python3 get.py macOS-Specific Notes:\nApple Silicon (M1/M2/M3): Toolchains are natively supported since ESP32 core 2.0.3+ If you get \u0026ldquo;permission denied\u0026rdquo; on serial port: sudo chmod 666 /dev/tty.usbserial-* macOS 10.15+ requires allowing \u0026ldquo;System Software from developer Espressif\u0026rdquo; in Security \u0026amp; Privacy during first upload Linux (Universal) # Ubuntu/Debian/Linux Mint # # 1. User permissions (dialout group) sudo usermod -aG dialout $USER # Log out and back in (or reboot) after this # 2. Install dependencies sudo apt-get update sudo apt-get install git python3-pip python3-pyserial python3-venv # 3. Create directories mkdir -p ~/Arduino/hardware/espressif cd ~/Arduino/hardware/espressif # 4. Clone and setup git clone --recursive https://github.com/espressif/arduino-esp32.git esp32 cd esp32 git checkout 3.0.7 # optional stable version cd tools python3 get.py Fedora/RHEL/CentOS/Rocky Linux # # 1. Groups (dialout on Fedora) sudo usermod -aG dialout $USER # 2. Dependencies sudo dnf install git python3-pip pyserial # 3. Setup directories and clone (same as Ubuntu) mkdir -p ~/Arduino/hardware/espressif cd ~/Arduino/hardware/espressif git clone --recursive https://github.com/espressif/arduino-esp32.git esp32 cd esp32/tools python3 get.py Arch Linux / Manjaro # # 1. Serial permissions (uucp/lock/tty, NOT dialout) sudo usermod -aG uucp,lock,tty $USER # CRITICAL: Log out and back in after this step # 2. Install dependencies sudo pacman -S git python-pip python-pyserial # 3. Clone ESP32 core mkdir -p ~/Arduino/hardware/espressif cd ~/Arduino/hardware/espressif git clone --recursive https://github.com/espressif/arduino-esp32.git esp32 cd esp32 git checkout 3.0.7 # optional: specific release cd tools python get.py openSUSE # # 1. Groups sudo usermod -aG dialout $USER # 2. Dependencies sudo zypper install git python3-pip python3-pyserial # 3. Clone (same structure as above) mkdir -p ~/Arduino/hardware/espressif cd ~/Arduino/hardware/espressif git clone --recursive https://github.com/espressif/arduino-esp32.git esp32 cd esp32/tools python3 get.py Platform-Specific Deep Configuration # Windows: Driver Installation # Most ESP32 boards use CP2102 or CH340 USB-to-Serial chips:\nCP2102 (Silicon Labs):\nDownload from: https://www.silabs.com/developers/usb-to-uart-bridge-vcp-drivers Install CP210x Universal Windows Driver Device Manager should show: \u0026ldquo;Silicon Labs CP210x USB to UART Bridge (COMx)\u0026rdquo; CH340/CH341 (WCH):\nWindows 10/11 usually auto-installs If not: https://www.wch.cn/downloads/CH341SER_ZIP.html Device Manager: \u0026ldquo;USB-SERIAL CH340 (COMx)\u0026rdquo; Verification:\n# In PowerShell Get-PnpDevice -Class Ports | Where-Object {$_.FriendlyName -like \u0026#34;*USB*\u0026#34;} macOS: Serial Port Permissions # Even after installation, you may get \u0026ldquo;Failed to connect to ESP32: Timed out waiting for packet header\u0026rdquo;:\n# Fix permissions (run each session or make permanent) sudo chmod 666 /dev/tty.usbserial-* # OR for specific port sudo chmod 666 /dev/cu.usbserial-0001 # Permanent fix: Create udev-like rule (macOS uses devfs) # Edit /etc/devfs.rules (create if doesn\u0026#39;t exist) sudo nano /etc/devfs.rules # Add: # [system=10] # add path \u0026#39;tty.usbserial*\u0026#39; mode 0666 # add path \u0026#39;cu.usbserial*\u0026#39; mode 0666 Linux: UDEV Rules (Permanent Fix) # Create persistent permissions so you don\u0026rsquo;t need sudo:\n# Create udev rule sudo nano /etc/udev/rules.d/50-esp32.rules Add these lines:\n# CP2102 (Espressif devkits) SUBSYSTEM==\u0026#34;usb\u0026#34;, ATTR{idVendor}==\u0026#34;10c4\u0026#34;, ATTR{idProduct}==\u0026#34;ea60\u0026#34;, MODE=\u0026#34;0666\u0026#34;, ENV{ID_MM_DEVICE_IGNORE}=\u0026#34;1\u0026#34;, ENV{ID_MM_PORT_IGNORE}=\u0026#34;1\u0026#34; # CH340/CH341 SUBSYSTEM==\u0026#34;usb\u0026#34;, ATTR{idVendor}==\u0026#34;1a86\u0026#34;, ATTR{idProduct}==\u0026#34;7523\u0026#34;, MODE=\u0026#34;0666\u0026#34;, ENV{ID_MM_DEVICE_IGNORE}=\u0026#34;1\u0026#34; # FT232 (some third-party boards) SUBSYSTEM==\u0026#34;usb\u0026#34;, ATTR{idVendor}==\u0026#34;0403\u0026#34;, ATTR{idProduct}==\u0026#34;6001\u0026#34;, MODE=\u0026#34;0666\u0026#34;, ENV{ID_MM_DEVICE_IGNORE}=\u0026#34;1\u0026#34; # ESP32-S2/S3 USB OTG (CDC) SUBSYSTEM==\u0026#34;usb\u0026#34;, ATTR{idVendor}==\u0026#34;303a\u0026#34;, MODE=\u0026#34;0666\u0026#34;, ENV{ID_MM_DEVICE_IGNORE}=\u0026#34;1\u0026#34; Reload rules:\nsudo udevadm control --reload-rules sudo udevadm trigger Verification \u0026amp; First Upload # 1. Select Correct Settings # Tools Menu:\nBoard: ESP32 Dev Module (or your specific model) Port: Windows: COM3 (or whatever appears when board is plugged in) macOS: /dev/cu.usbserial-0001 or /dev/tty.usbserial-* Linux: /dev/ttyUSB0 or /dev/ttyACM0 Upload Speed: 921600 (or 115200 if unreliable) CPU Frequency: 240MHz (WiFi/BT) Flash Frequency: 80MHz Flash Mode: QIO Partition Scheme: Default 4MB with spiffs (or Huge APP if large sketches) 2. Test Sketch # // Built-in example: File → Examples → 01.Basics → Blink // Modify LED_BUILTIN to 2 (most ESP32 devkits use GPIO2 for onboard LED) #define LED_BUILTIN 2 void setup() { pinMode(LED_BUILTIN, OUTPUT); Serial.begin(115200); } void loop() { digitalWrite(LED_BUILTIN, HIGH); Serial.println(\u0026#34;LED ON\u0026#34;); delay(1000); digitalWrite(LED_BUILTIN, LOW); Serial.println(\u0026#34;LED OFF\u0026#34;); delay(1000); } 3. Upload Process # Hold BOOT button on ESP32 (if your board has one) Click Upload in IDE Release BOOT button when you see \u0026ldquo;Connecting\u0026hellip;\u0026rdquo; in the console If successful: \u0026ldquo;Hash of data verified\u0026rdquo; followed by \u0026ldquo;Hard resetting via RTS pin\u0026hellip;\u0026rdquo; Auto-reset issues: If upload fails consistently, you may need a 10µF capacitor between EN (RST) and GND, or use specific reset methods (varies by board design).\nTroubleshooting Matrix # Symptom Windows macOS Linux Port not visible Install drivers (CP2102/CH340) Install drivers via brew or manually Add user to dialout/uucp group Permission denied Run IDE as Administrator (temporary) chmod 666 /dev/tty.* Udev rules or sudo usermod -aG dialout $USER Connection timed out Wrong COM port, try lower baud (115200) Check cable, try /dev/cu.* instead of /dev/tty.* Check dmesg | grep tty for port name Failed to write to target RAM USB 3.0 issue, try USB 2.0 hub USB-C to USB-A adapter issue Add user to lock group (Arch) Missing Python/serial pip install pyserial pip3 install pyserial sudo pacman -S python-pyserial (Arch) or apt install python3-serial Specific Error Messages # \u0026ldquo;A fatal error occurred: Failed to connect to ESP32: Timed out\u0026hellip;\u0026rdquo;\nCause: ESP32 not in bootloader mode Fix: Hold BOOT button, click Upload, release when \u0026ldquo;Connecting\u0026hellip;\u0026rdquo; appears Permanent Fix: Some boards need DTR/RTS capacitors soldered for auto-reset \u0026ldquo;Permission denied: \u0026lsquo;/dev/ttyUSB0\u0026rsquo;\u0026rdquo;\nLinux: sudo usermod -aG dialout $USER then reboot (not just logout) Arch: Use uucp group instead of dialout \u0026ldquo;pyserial or esptool not found\u0026rdquo;\n# All platforms pip3 install --user pyserial esptool # Or for system-wide (Linux) sudo pip3 install pyserial esptool \u0026ldquo;Cannot open /dev/ttyUSB0: Device or resource busy\u0026rdquo;\nCause: ModemManager is using the port Fix: sudo systemctl stop ModemManager (Linux) or add udev rule with ENV{ID_MM_DEVICE_IGNORE}=\u0026quot;1\u0026quot; Advanced Topics # Version Management (Manual Method) # cd ~/Arduino/hardware/espressif/esp32 # List available versions git tag | grep -E \u0026#34;^3\\.|^2\\.\u0026#34; | tail -20 # Switch to specific version git checkout 3.0.7 git submodule update --init --recursive # Critical step cd tools \u0026amp;\u0026amp; python get.py # Return to bleeding edge git checkout master git pull git submodule update --init --recursive python get.py Offline Installation # For air-gapped systems:\nOn connected machine: Clone repo, run get.py Copy entire ~/Arduino/hardware/espressif/esp32 folder to offline machine Install Arduino IDE on offline machine Paste folder to same location Multiple Arduino Versions # If using IDE 1.8 and 2.x simultaneously:\nIDE 1.8: Uses ~/Arduino/hardware/ (Linux/mac) or Documents\\Arduino\\hardware\\ (Windows) IDE 2.x: Uses ~/.arduino15/packages/ for Boards Manager installs Manual method works for both if placed in correct location ESP32-S2/S3/C3 Specifics # USB-OTG Native (S2/S3):\nNo drivers needed on Windows 10/11, macOS, Linux Port appears as USB JTAG/serial debug unit Hold BOOT, press RST (on board), release BOOT to enter bootloader C3 (RISC-V):\nUses different toolchain (RISC-V instead of Xtensa) Automatically handled by core 2.0.0+ Maintenance # Updating (Boards Manager) # Tools → Board → Boards Manager → Updates (checkmark icon) → Update ESP32\nUpdating (Manual) # cd ~/Arduino/hardware/espressif/esp32 git pull git submodule update --init --recursive cd tools python get.py Complete Removal # Boards Manager: Boards Manager → ESP32 → Remove Manual: rm -rf ~/Arduino/hardware/espressif/esp32\nFinal Checklist:\nArduino IDE installed and running Board URL added (for Boards Manager) OR Git repo cloned User added to correct serial group (Linux) OR Drivers installed (Windows) Board selected matching your hardware Correct port selected USB cable is data-capable (test with phone file transfer) Bootloader mode activated (if auto-reset fails) This guide covers 95% of installation scenarios. For specific board variants (LOLIN, M5Stack, etc.), consult the manufacturer\u0026rsquo;s additional instructions after completing this base installation.\nFAQ: Why \u0026amp; When to Choose Manual Installation # 1. Why should I install the ESP32 core manually instead of using Boards Manager? # Boards Manager downloads ~800MB–1.2GB of toolchain binaries (xtensa compilers, esptool, precompiled libraries) through Arduino IDE\u0026rsquo;s package manager. This process is atomic — if it fails at 99%, it restarts from zero.\nManual installation gives you:\nResumable downloads: git clone supports resume (git fetch --depth=1), and get.py can be re-run until completion ** granular control**: Download toolchains once, copy to multiple machines Offline capability: Works on air-gapped systems after initial setup Bleeding-edge fixes: Access commits newer than the latest \u0026ldquo;stable\u0026rdquo; release (e.g., urgent WiFi patches not yet in 3.0.7) Choose Manual if: Your connection drops frequently, you manage multiple dev machines, or you need unreleased bug fixes.\n2. I have a weak internet connection and keep getting \u0026ldquo;Download timeout\u0026rdquo; or \u0026ldquo;CRC check failed\u0026rdquo; errors. How does manual installation solve this? # The Problem: Arduino Boards Manager uses Java\u0026rsquo;s network stack with fixed 30–60 second timeouts. On slow connections (\u0026lt;512kbps) or high-latency networks (satellite, rural DSL), the toolchain download (particularly the 500MB+ GCC compiler package) times out repeatedly.\nThe Manual Solution:\n# Step 1: Shallow clone (only ~150MB of metadata, not full history) git clone --depth=1 --recursive https://github.com/espressif/arduino-esp32.git esp32 # Step 2: Run get.py with resume capability cd esp32/tools python get.py # If it fails at 45%, simply run python get.py again — it resumes partial files Pro tip for unstable connections: Use wget or a download manager for the toolchains manually, then place them in esp32/tools/dist/ before running get.py (which will detect existing files and skip re-downloading).\n3. Can I resume an interrupted installation, or do I have to start over? # Boards Manager: No. If the progress bar freezes at 70% and you cancel, the partial cache is discarded. Next attempt restarts from 0%.\nManual Method: Yes, inherently resumable:\nGit: git fetch --unshallow or git submodule update --init resumes interrupted clones Toolchains: get.py checks SHA256 hashes of files in tools/dist/. Completed files are skipped; incomplete files are re-downloaded. Recovery procedure after disconnect:\ncd ~/Arduino/hardware/espressif/esp32 git submodule update --init --recursive --progress # Shows % complete cd tools \u0026amp;\u0026amp; python get.py --verbose # Shows which specific tool is being fetched 4. How do I install the ESP32 core on a computer with no internet connection (air-gapped)? # Boards Manager: Impossible. It requires live JSON index validation and real-time package download.\nManual Method: The \u0026ldquo;Sneakernet\u0026rdquo; approach:\nOn connected machine: mkdir -p ~/Arduino/hardware/espressif cd ~/Arduino/hardware/espressif git clone --recursive https://github.com/espressif/arduino-esp32.git esp32 cd esp32/tools \u0026amp;\u0026amp; python get.py cd ../../.. tar -czf esp32-offline.tar.gz espressif/ # Compress entire core + toolchains Transfer esp32-offline.tar.gz via USB drive to offline machine On offline machine: mkdir -p ~/Arduino/hardware/ tar -xzf esp32-offline.tar.gz -C ~/Arduino/hardware/ Install Python dependencies manually (download .whl files from PyPI on connected machine, transfer via USB): pip install --no-index --find-links=/path/to/usb/pyserial pyserial Result: Full ESP32 support without the offline machine ever touching the internet.\n5. The Boards Manager download is huge (1GB+) and fails repeatedly on my metered connection. Is manual installation smaller? # Size comparison:\nBoards Manager: Downloads all available toolchains (xtensa-esp32, xtensa-esp32s2, xtensa-esp32s3, riscv32-esp32c3, esp32ulp) plus all precompiled libraries — ~1.2GB Manual (selective): You can manually download only the toolchain for your specific chip Bandwidth-saving manual approach:\n# After cloning, edit esp32/tools/get.py to comment out unneeded platforms # Or manually download only your chip\u0026#39;s toolchain from: # https://github.com/espressif/crosstool-NG/releases # Place in tools/dist/ then run get.py (it will use local files) Typical savings: If you only use ESP32-WROOM (Xtensa), you can skip RISC-V and ULP tools, reducing download to ~400MB.\n6. Does manual installation make it easier to manage multiple developer machines? # Yes — significantly.\nThe \u0026ldquo;Golden Master\u0026rdquo; Method:\nInstall manually on Machine A (with good internet) Verify it compiles and uploads correctly Archive the entire ~/Arduino/hardware/espressif/esp32 folder Distribute to Machines B, C, D via local network or USB drives No per-machine downloads: Each developer doesn\u0026rsquo;t need to download 1GB through the company firewall.\nVersion consistency: All machines use the exact same git commit (e.g., 3.0.7), preventing \u0026ldquo;works on my machine\u0026rdquo; issues caused by Boards Manager auto-updating on one laptop but not another.\n7. Why does get.py fail with \u0026ldquo;Connection timeout\u0026rdquo; even when I can browse websites fine? # Root causes:\nParallel downloads: get.py attempts to download 4–6 large toolchain archives simultaneously. On weak connections, this saturates the pipe and causes timeouts. CDN issues: GitHub releases (where toolchains are hosted) throttle connections from certain regions or ISPs. SSL inspection: Corporate firewalls intercept HTTPS, breaking Python\u0026rsquo;s SSL verification. Fixes for weak connections:\n# Single-threaded download (slower but stable) python get.py --platform esp32 # Only download ESP32 tools, not S2/S3/C3 # If behind corporate proxy export HTTP_PROXY=http://proxy.company.com:8080 export HTTPS_PROXY=http://proxy.company.com:8080 python get.py # Disable SSL verify (last resort, security risk) python get.py --ssl-no-verify 8. Is manual installation more stable than Boards Manager for daily development? # Stability factors:\nAspect Boards Manager Manual/Git Update surprise Auto-updates can break sketches overnight You control when to git pull Network dependency Requires internet to verify packages every launch Fully offline after setup Corruption recovery Delete entire package, re-download 1GB git status shows exactly which file is corrupted Rollback Difficult (must uninstall, reinstall old version) git checkout 2.0.17 instant rollback Trade-off: Manual requires you to remember to update (git pull), while Boards Manager auto-updates. For weak internet, manual is more stable because it won\u0026rsquo;t randomly fail to compile because it can\u0026rsquo;t reach the package index server.\n9. How do I update the ESP32 core manually when I have limited monthly bandwidth? # Incremental updates (Manual advantage):\ncd ~/Arduino/hardware/espressif/esp32 # Check what\u0026#39;s new without downloading git fetch origin git log HEAD..origin/master --oneline --stat # If only small changes (e.g., 3 files changed, +50 lines): git pull origin master # Toolchains usually don\u0026#39;t change between minor commits, so get.py does nothing # If major release (e.g., 3.0.7 → 3.1.0): # Only then run get.py to fetch new toolchains if necessary python tools/get.py Bandwidth conservation: Git pulls are ~KBs of text. Boards Manager re-downloads entire toolchain packages even for tiny core updates.\nSelective updates: You can update only the libraries/ folder for bug fixes without touching the massive tools/ folder.\n10. Can I mix manual and Boards Manager installations on the same computer? # Yes, but with caution.\nPriority rules:\nArduino IDE prioritizes Manual (~/Arduino/hardware/) over Boards Manager (~/.arduino15/packages/) If both exist, the manual version is used for compilation Use case — \u0026ldquo;Hybrid Mode\u0026rdquo;:\nInstall stable via Boards Manager: For production work (reliable, no internet needed after install) Install dev version manually: For testing new features Switch between them: Rename ~/Arduino/hardware/espressif to ~/Arduino/hardware/espressif_dev when you want Boards Manager stable Rename back when you need bleeding edge Conflict warning: Never have the same version in both locations (e.g., 3.0.7 in Boards Manager and 3.0.7 in manual). This causes \u0026ldquo;Multiple libraries found\u0026rdquo; warnings and undefined behavior.\nBest practice: Use Boards Manager for stable work, manual only for specific development/testing — or uninstall Boards Manager version entirely to save disk space once manual is working.\nQuick Decision Matrix # Your Situation Recommended Method DSL/Satellite/Spotty 4G Manual — resume capability essential Corporate firewall/proxy Manual — easier to configure proxy settings Single machine, good internet Boards Manager — simpler updates Air-gapped/Offline lab Manual — only viable option Teaching/classroom (30 students) Manual — install once, copy to 30 USB drives Need yesterday\u0026rsquo;s bugfix Manual — access commits immediately, no wait for release Bottom line: Manual installation transforms the ESP32 core from a \u0026ldquo;streaming service\u0026rdquo; (requires constant reliable internet) into a \u0026ldquo;downloaded application\u0026rdquo; (works offline, survives interruptions).\n","date":"31 March 2026","externalUrl":null,"permalink":"/tutorials-guides/esp32-arduino-driver/","section":"Tutorials \u0026 Technical Guides","summary":" Comprehensive ESP32 Arduino Core Installation Guide # Overview: Two Installation Methods # Method Best For Pros Cons Boards Manager Beginners, stability One-click install, version management, automatic updates Slight delay for newest features Manual/Git Clone Developers, beta testers Latest commits, instant updates, offline capable Manual updates, potential instability Method 1: Boards Manager (Recommended for Most Users) # Universal Prerequisites # Arduino IDE 1.8.19+ or Arduino IDE 2.3+ (IDE 2.x recommended) Python 3.7+ (usually bundled with IDE 2.x) USB Cable: Must be data-capable, not charge-only Step 1: Add ESP32 Board Manager URL # Arduino IDE 2.x:\nFile → Preferences (or Ctrl+,) In “Additional boards manager URLs” add: https://raw.githubusercontent.com/espressif/arduino-esp32/gh-pages/package_esp32_index.json Click OK Arduino IDE 1.8.x:\nFile → Preferences Same URL as above If multiple URLs, separate with commas Step 2: Install the Core # Tools → Board → Boards Manager… Search for “ESP32” by Espressif Systems Select version (latest stable recommended, e.g., 3.0.x or 2.0.x) Click Install (downloads ~500MB-1GB of toolchains) Step 3: Select Your Board # Tools → Board → ESP32 Arduino → Select your specific board:\n","title":"ESP32 Arduino Core Installation Guide","type":"tutorials-guides"},{"content":" Markdown‑First Workflow (for Engineers) # This is a professional, Markdown‑first authoring workflow for technical documentation, lectures, blogs, and slides.\nAll content is written in Markdown, and outputs (PDF, HTML, slides, static site) are generated automatically.\n1. Core tools # VS Code – main editor. Markdown – single source format for docs, lectures, blog posts. LaTeX – inline and block math inside Markdown. Mermaid – diagrams inside code blocks. Marp – Markdown‑based slides. Hugo – static site generator for SEO‑friendly blog / docs. Pandoc + markdown-pdf – PDF / HTML / DOCX generation from Markdown. 2. Writing layer (what you type) # All content is written in Markdown:\n$$ \\ddot{x} + 2\\zeta\\omega_n\\dot{x} + \\omega_n^2 x = 0 $$ \\(\\ddot{x} + 2\\zeta\\omega_n\\dot{x} + \\omega_n^2 x = 0\\)\nDiagrams:\ngraph LR A[FIXED] --\u0026gt; B[Mermaid Works] graph LR A[FIXED] --\u003e B[Mermaid Works] Slides (Marp):\n--- marp: true theme: default paginate: true math: mathjax --- # Lecture 1 – Introduction to Robotics 3. Editor setup (VS Code) # Install these (or similar):\nMarp extension – live preview + export of slides. Mermaid extensions Mermaid Chart or Mermaid Graphical Editor vscode-mermaid-preview / markdown-mermaid Latex‑related LaTeX Workshop Pandoc + PDF vscode-pandoc markdown-pdf Markdown enhancements markdown-all-in-one markdownlint You can:\nEdit Markdown + LaTeX + Mermaid in one pane. Open a Mermaid‑editing webview attached to any mermaid block. Preview Mermaid and math in the editor or side panel. 4. Outputs (what gets published) # From the same Markdown, generate:\nStatic site / blog Hugo builds SEO‑friendly static pages from Markdown + LaTeX + Mermaid. PDFs markdown-pdf → quick Chromium‑style PDFs. pandoc + LaTeX backend → high‑quality PDFs (for reports, notes, books). DOCX / other formats pandoc can export to DOCX, EPUB, HTML, etc. Slides Marp exports the same Markdown file as PDF / HTML slides. 5. Folder structure (example) # project/ ├── docs/ # docs, guides, tutorials ├── lectures/ # Markdown lecture notes ├── slides/ # Marp decks ├── blog/ # Hugo content (or Hugo subdir) ├── pdf/ # output PDFs ├── html/ # output HTML versions └── .vscode/ # workspace settings + tasks Git‑native, repo‑first. All source is Markdown + LaTeX + Mermaid. Generated outputs are not hand‑edited. 6. Bilingual + SEO (Hugo) # Use Hugo with en / ar (or other languages): content/en/posts/... content/ar/posts/... In Hugo templates, add: canonical URL hreflang=\u0026quot;en\u0026quot; / hreflang=\u0026quot;ar\u0026quot; This gives you SEO‑friendly static pages with proper language signals. 7. One‑command workflow # You can automate with just / make / scripts:\njust hugo # build Hugo site just pdf file.md # build PDF from Markdown just slideshow.md # export Marp deck to PDF Plus CI/CD (e.g., GitHub Actions) that:\nBuilds Hugo site Runs pandoc / markdown-pdf Deploys static site to GitHub Pages / Netlify. 8. Key philosophy # One source format: Markdown + LaTeX + Mermaid. Many outputs: HTML, PDF, DOCX, slides, static site. Git‑first: all content is in version‑controlled Markdown. Tool‑agnostic style: you can swap Hugo ↔ others later; Markdown stays. ","date":"31 March 2026","externalUrl":null,"permalink":"/tutorials-guides/markdown-workflow/","section":"Tutorials \u0026 Technical Guides","summary":"Markdown‑First Workflow (for Engineers) # This is a professional, Markdown‑first authoring workflow for technical documentation, lectures, blogs, and slides.\nAll content is written in Markdown, and outputs (PDF, HTML, slides, static site) are generated automatically.\n1. Core tools # VS Code – main editor. Markdown – single source format for docs, lectures, blog posts. LaTeX – inline and block math inside Markdown. Mermaid – diagrams inside code blocks. Marp – Markdown‑based slides. Hugo – static site generator for SEO‑friendly blog / docs. Pandoc + markdown-pdf – PDF / HTML / DOCX generation from Markdown. 2. Writing layer (what you type) # All content is written in Markdown:\n$$ \\ddot{x} + 2\\zeta\\omega_n\\dot{x} + \\omega_n^2 x = 0 $$ \\(\\ddot{x} + 2\\zeta\\omega_n\\dot{x} + \\omega_n^2 x = 0\\)\nDiagrams:\ngraph LR A[FIXED] --\u003e B[Mermaid Works] graph LR A[FIXED] --\u003e B[Mermaid Works] Slides (Marp):\n--- marp: true theme: default paginate: true math: mathjax --- # Lecture 1 – Introduction to Robotics 3. Editor setup (VS Code) # Install these (or similar):\n","title":"Markdown‑First Workflow (for Engineers)","type":"tutorials-guides"},{"content":" Mentorship \u0026amp; Consultations # خدمات الإرشاد والاستشارات # هذه الصفحة مخصصة لخدمات الإرشاد (Mentorship) والاستشارات (Consultations)، لزيارة صفحة الكورسات.\nالخدمات المتاحة # الإرشاد الأكاديمي والمنح\nUWC/Bachelor scholarships + خطة تجهيز ملف التقديم. الإرشاد المهني للهوية المهنية\nProfessional personal branding + career positioning. جلسة استشارة تقنية\nجلسة استشارية بموضوع تختاره أنت من نطاق خبراتي العملية. سياسة الحجز والدفع # وسيلة الدفع: تحويل شام كاش (ShamCash). تأكيد الحجز يتطلب دفعًا كاملًا مسبقًا. أي حجز بدون تأكيد الدفع يتم إلغاؤه تلقائيًا. يسمح بإعادة جدولة مرة واحدة فقط عند إشعار مسبق 24 ساعة. لا يوجد استرداد بعد تنفيذ الجلسة. نظام التسعير الذكي (حسب الجهة الجغرافية) # داخل سوريا: 25$ للجلسة. المغتربون/الخليج: 50$ إلى 75$ للجلسة (حسب نوع الطلب وتعقيده). دولي: 100$ للجلسة. International Payments (Quick Note) # للتحويلات الدولية، تواصل معي عبر البريد: contact@mulhamfetna.com\nFor international payments, contact me directly at: contact@mulhamfetna.com\nتنبيه مهم: للمدفوعات الدولية، تواصل معي مباشرة على contact@mulhamfetna.com وسيتم التعامل يدويًا مؤقتًا إلى أن يصبح حجم الطلب مبررًا لنظام دفع آلي.\nالإبلاغ عن المشاكل # في حال وجود أي مشكلة في الحجز أو الدفع، أرسل التفاصيل مع لقطات شاشة إلى:\ncontact@mulhamfetna.com\nEnglish Summary # This page is for mentorship and consultation services, to visit my courses page.\nUse the service pages above to choose the right offer for your goal.\nBooking \u0026amp; Payment Policy # Payment method: ShamCash transfer. Full prepayment is required to confirm a booking. Any unpaid booking is automatically cancelled. One reschedule is allowed with 24-hour notice. No refund after session delivery. Smart Pricing Model (Geo-based) # Inside Syria: $25/session. Diaspora/Gulf: $50–$75/session (based on request type and complexity). International: $100/session. Important: For international payments, contact me directly at contact@mulhamfetna.com and payment will be handled manually until volume justifies a dedicated system.\nSupport # For booking/payment issues, email details and screenshots to:\ncontact@mulhamfetna.com\n","date":"31 March 2026","externalUrl":null,"permalink":"/services/","section":"Mentorship \u0026 Consultations","summary":"Mentorship \u0026 Consultations # خدمات الإرشاد والاستشارات # هذه الصفحة مخصصة لخدمات الإرشاد (Mentorship) والاستشارات (Consultations)، لزيارة صفحة الكورسات.\nالخدمات المتاحة # الإرشاد الأكاديمي والمنح\nUWC/Bachelor scholarships + خطة تجهيز ملف التقديم. الإرشاد المهني للهوية المهنية\nProfessional personal branding + career positioning. جلسة استشارة تقنية\nجلسة استشارية بموضوع تختاره أنت من نطاق خبراتي العملية. سياسة الحجز والدفع # وسيلة الدفع: تحويل شام كاش (ShamCash). تأكيد الحجز يتطلب دفعًا كاملًا مسبقًا. أي حجز بدون تأكيد الدفع يتم إلغاؤه تلقائيًا. يسمح بإعادة جدولة مرة واحدة فقط عند إشعار مسبق 24 ساعة. لا يوجد استرداد بعد تنفيذ الجلسة. نظام التسعير الذكي (حسب الجهة الجغرافية) # داخل سوريا: 25$ للجلسة. المغتربون/الخليج: 50$ إلى 75$ للجلسة (حسب نوع الطلب وتعقيده). دولي: 100$ للجلسة. International Payments (Quick Note) # للتحويلات الدولية، تواصل معي عبر البريد: contact@mulhamfetna.com\n","title":"Mentorship \u0026 Consultations","type":"services"},{"content":" الإرشاد المهني وبناء الهوية المهنية # لمن هذه الخدمة؟ # هذه الخدمة مناسبة لك إذا:\nتملك مهارات أو إنجازات لكن لا تعرف كيف تعرضها بشكل مقنع. تشعر بالتشتت بين مجالات عديدة وتحتاج مسارًا مهنيًا واضحًا. تريد تحسين صورتك المهنية للفرص الأكاديمية أو الوظيفية أو المستقلة. ماذا ستحصل؟ # وضوح في اتجاهك المهني والأولويات الفعلية. صياغة احترافية لهويتك المهنية وإنجازاتك. خطة تنفيذ عملية (أنشطة + مهارات + أدوات + شهادات مناسبة). مراجعة دورية للمخرجات (ضمن البرنامج الكامل). منهجية الجلسات (3 مراحل) # المرحلة 1: التحليل والعصف الذهني # استخراج خبراتك السابقة وتحليل نقاط القوة والفجوات. تحديد نقطة التركيز الأكثر تأثيرًا الآن. المرحلة 2: تصميم المسار # بناء خطة مهنية واضحة ومناسبة لشخصيتك ومؤهلاتك. تحويل التشتت إلى roadmap قابلة للتنفيذ. المرحلة 3: الأدوات والتنفيذ # تحديد الأدوات والمهارات والشهادات الأكثر قيمة لحالتك. ترتيب خطوات تطبيق أسبوعية واضحة. الصيغ والأسعار # التسعير حسب الجهة الجغرافية # داخل سوريا: 25$ للجلسة. المغتربون/الخليج: 50$ إلى 75$ للجلسة. دولي: 100$ للجلسة. ملاحظة عن البرنامج الكامل # تسعير البرنامج الكامل (جلسات + متابعة) يتم تحديده بعد جلسة التشخيص حسب حاجتك.\nسياسة الحجز والدفع # الدفع عبر ShamCash فقط. تأكيد الحجز يتطلب دفعًا كاملًا مسبقًا. أي حجز بدون تأكيد الدفع يتم إلغاؤه تلقائيًا. يسمح بإعادة جدولة مرة واحدة عند إشعار مسبق 24 ساعة. لا يوجد استرداد بعد تنفيذ الجلسة. تنبيه مهم: للمدفوعات الدولية، تواصل معي مباشرة على contact@mulhamfetna.com وسيتم التعامل يدويًا مؤقتًا إلى أن يصبح حجم الطلب مبررًا لنظام دفع آلي.\nخطوات الحجز (مهم) # حوّل الرسوم عبر ShamCash أولًا. احجز الموعد عبر Cal. أي حجز بدون تأكيد الدفع يتم إلغاؤه تلقائيًا. معلومات الدفع عبر ShamCash # Payment Identifier: d9baec7d91b344eb9c2d04363dd84664 للموبايل (مسح مباشر): رابط الحجز المباشر # Book your 1-hour meeting\nالنتيجة المتوقعة # الهدف ليس فقط الحصول على فرصة قصيرة المدى، بل بناء هوية مهنية واضحة ومستدامة يمكن أن تنمو معها فرصك على المدى الطويل.\nالدعم وحل المشاكل # لأي مشكلة في الحجز أو الدفع، أرسل التفاصيل مع لقطات شاشة إلى: contact@mulhamfetna.com\nBooking Widget # English Snapshot # Professional mentorship for career positioning and personal brand clarity, including a practical execution roadmap and optional ongoing follow-up.\nBooking \u0026amp; Payment # Pay first via ShamCash (Identifier: d9baec7d91b344eb9c2d04363dd84664). Direct booking link: https://cal.com/mulham-fetna/1-hour-meeting Pricing: $25 (Syria), $50–$75 (Diaspora/Gulf), $100 (International) per session. For international payments, contact: contact@mulhamfetna.com for manual handling. ","date":"31 March 2026","externalUrl":null,"permalink":"/services/professional-mentorship/","section":"Mentorship \u0026 Consultations","summary":" الإرشاد المهني وبناء الهوية المهنية # لمن هذه الخدمة؟ # هذه الخدمة مناسبة لك إذا:\nتملك مهارات أو إنجازات لكن لا تعرف كيف تعرضها بشكل مقنع. تشعر بالتشتت بين مجالات عديدة وتحتاج مسارًا مهنيًا واضحًا. تريد تحسين صورتك المهنية للفرص الأكاديمية أو الوظيفية أو المستقلة. ماذا ستحصل؟ # وضوح في اتجاهك المهني والأولويات الفعلية. صياغة احترافية لهويتك المهنية وإنجازاتك. خطة تنفيذ عملية (أنشطة + مهارات + أدوات + شهادات مناسبة). مراجعة دورية للمخرجات (ضمن البرنامج الكامل). منهجية الجلسات (3 مراحل) # المرحلة 1: التحليل والعصف الذهني # استخراج خبراتك السابقة وتحليل نقاط القوة والفجوات. تحديد نقطة التركيز الأكثر تأثيرًا الآن. المرحلة 2: تصميم المسار # بناء خطة مهنية واضحة ومناسبة لشخصيتك ومؤهلاتك. تحويل التشتت إلى roadmap قابلة للتنفيذ. المرحلة 3: الأدوات والتنفيذ # تحديد الأدوات والمهارات والشهادات الأكثر قيمة لحالتك. ترتيب خطوات تطبيق أسبوعية واضحة. الصيغ والأسعار # التسعير حسب الجهة الجغرافية # داخل سوريا: 25$ للجلسة. المغتربون/الخليج: 50$ إلى 75$ للجلسة. دولي: 100$ للجلسة. ملاحظة عن البرنامج الكامل # تسعير البرنامج الكامل (جلسات + متابعة) يتم تحديده بعد جلسة التشخيص حسب حاجتك.\n","title":"Professional Personal Brand Mentorship","type":"services"},{"content":" الإرشاد الأكاديمي والمنح (UWC / Bachelor Scholarships) # لمن هذه الخدمة؟ # هذه الخدمة مناسبة لك إذا:\nتخطط للتقديم على منح UWC أو منح البكالوريوس الدولية. تشعر بالتشتت بين المتطلبات (مقالات، أنشطة، اختبارات، timeline). تريد تحويل إنجازاتك الحالية إلى ملف تقديم تنافسي ومنطقي. ماذا ستحصل؟ # تقييم واضح لوضعك الحالي مقارنة بمتطلبات المنح المستهدفة. صياغة احترافية لإنجازاتك وخبراتك بصيغة مناسبة للأبلكيشن. خطة أنشطة واقعية (قصيرة/متوسطة المدى) تدعم فرص القبول. مراجعة دورية للمقالات والتقديمات ضمن البرنامج الكامل. آلية العمل # المرحلة 1: التشخيص والتوجيه # جلسة عصف ذهني لاستخراج كل خبراتك وإنجازاتك. تحديد الفجوات الحقيقية في ملفك مقارنة بالهدف. المرحلة 2: استراتيجية التقديم # اختيار المسار المناسب (تقديم الآن / تجهيز مرحلي). بناء narrative واضح لهويتك الأكاديمية والشخصية. المرحلة 3: التنفيذ والمراجعة # مراجعة المقالات والتقديمات. متابعة القرار النهائي حتى لحظة Submit (ضمن البرنامج الكامل). المدة المتوقعة # غالبًا: 3 إلى 5 جلسات حسب الحالة. المتابعة الدورية متاحة في البرنامج الكامل. الأسعار # التسعير حسب الجهة الجغرافية # داخل سوريا: 25$ للجلسة. المغتربون/الخليج: 50$ إلى 75$ للجلسة. دولي: 100$ للجلسة. ملاحظة عن البرنامج الكامل # تسعير البرنامج الكامل (3–5 جلسات + متابعة) يتم تحديده بعد جلسة التشخيص حسب الحاجة الفعلية.\nسياسة الحجز والدفع # الدفع عبر ShamCash فقط. تأكيد الحجز يتطلب دفعًا كاملًا مسبقًا. أي حجز بدون تأكيد الدفع يتم إلغاؤه تلقائيًا. يسمح بإعادة جدولة مرة واحدة عند إشعار مسبق 24 ساعة. لا يوجد استرداد بعد تنفيذ الجلسة. تنبيه مهم: للمدفوعات الدولية، تواصل معي مباشرة على contact@mulhamfetna.com وسيتم التعامل يدويًا مؤقتًا إلى أن يصبح حجم الطلب مبررًا لنظام دفع آلي.\nخطوات الحجز (مهم) # حوّل الرسوم عبر ShamCash أولًا. احجز الموعد عبر Cal. أي حجز بدون تأكيد الدفع يتم إلغاؤه تلقائيًا. معلومات الدفع عبر ShamCash # Payment Identifier: d9baec7d91b344eb9c2d04363dd84664 للموبايل (مسح مباشر): رابط الحجز المباشر # Book your 1-hour meeting\nملاحظات مهمة # أنا لا أقدم بدلًا عنك؛ أنا أدربك لتقديم نفسك بأفضل صورة. الهدف ليس فقط التقديم، بل بناء خطة أكاديمية ذكية قابلة للتنفيذ. الدعم وحل المشاكل # لأي مشكلة في الحجز أو الدفع، أرسل التفاصيل مع لقطات شاشة إلى: contact@mulhamfetna.com\nBooking Widget # English Snapshot # Scholarships mentorship focused on UWC/Bachelor applications: profile positioning, application strategy, essay review, and submission readiness with clear milestones.\nBooking \u0026amp; Payment # Pay first via ShamCash (Identifier: d9baec7d91b344eb9c2d04363dd84664). Direct booking link: https://cal.com/mulham-fetna/1-hour-meeting Pricing: $25 (Syria), $50–$75 (Diaspora/Gulf), $100 (International) per session. For international payments, contact: contact@mulhamfetna.com for manual handling. ","date":"31 March 2026","externalUrl":null,"permalink":"/services/academic-mentorship/","section":"Mentorship \u0026 Consultations","summary":" الإرشاد الأكاديمي والمنح (UWC / Bachelor Scholarships) # لمن هذه الخدمة؟ # هذه الخدمة مناسبة لك إذا:\nتخطط للتقديم على منح UWC أو منح البكالوريوس الدولية. تشعر بالتشتت بين المتطلبات (مقالات، أنشطة، اختبارات، timeline). تريد تحويل إنجازاتك الحالية إلى ملف تقديم تنافسي ومنطقي. ماذا ستحصل؟ # تقييم واضح لوضعك الحالي مقارنة بمتطلبات المنح المستهدفة. صياغة احترافية لإنجازاتك وخبراتك بصيغة مناسبة للأبلكيشن. خطة أنشطة واقعية (قصيرة/متوسطة المدى) تدعم فرص القبول. مراجعة دورية للمقالات والتقديمات ضمن البرنامج الكامل. آلية العمل # المرحلة 1: التشخيص والتوجيه # جلسة عصف ذهني لاستخراج كل خبراتك وإنجازاتك. تحديد الفجوات الحقيقية في ملفك مقارنة بالهدف. المرحلة 2: استراتيجية التقديم # اختيار المسار المناسب (تقديم الآن / تجهيز مرحلي). بناء narrative واضح لهويتك الأكاديمية والشخصية. المرحلة 3: التنفيذ والمراجعة # مراجعة المقالات والتقديمات. متابعة القرار النهائي حتى لحظة Submit (ضمن البرنامج الكامل). المدة المتوقعة # غالبًا: 3 إلى 5 جلسات حسب الحالة. المتابعة الدورية متاحة في البرنامج الكامل. الأسعار # التسعير حسب الجهة الجغرافية # داخل سوريا: 25$ للجلسة. المغتربون/الخليج: 50$ إلى 75$ للجلسة. دولي: 100$ للجلسة. ملاحظة عن البرنامج الكامل # تسعير البرنامج الكامل (3–5 جلسات + متابعة) يتم تحديده بعد جلسة التشخيص حسب الحاجة الفعلية.\n","title":"Scholarships Mentorship","type":"services"},{"content":" Mulham Fetna # Mechatronics \u0026amp; AI/ML Engineer · Researcher · Technical Trainer\nI build robotics and machine-learning systems, publish research, and train the next generation of Syrian engineers and researchers — from Aleppo.\nMechatronics engineering student at the University of Aleppo (fourth year, top of class at 89%). Open-source contributor with 3 pull requests merged into OpenCV and OpenDR — two of them into OpenCV core.\nTeaching since 2022 across Syria, Lebanon, Palestine, and internationally through Paper Airplanes and HerWill — 500\u0026#43; learners across all programs, 2022–2026.\nThis site is my platform-independent professional record — built after losing access to LinkedIn, and designed to outlast any single platform.\nThe mission # I work at the intersection of things that are usually kept apart: mechatronics, data science, MLOps, research, STEM training, and open source.\nSyria\u0026rsquo;s engineering talent is not the bottleneck — access is. Access to a research culture, to global open-source communities, to the tools and the credentials that turn ability into opportunity. Everything I build is aimed at that gap, and at making the path I had to improvise into one that others can simply follow.\nReach: Syria → the Arab world → globally.\nExperience at a glance: 4 years teaching · 500\u0026#43; learners across all programs, 2022–2026 · 3 merged PRs into OpenCV and OpenDR · 2 companies founded\nWhat I do # I founded and lead two Syrian ventures — Neurobotics (engineering and technical education) and Boundless (academic services, partnered with SANAD and UNFPA Syria). Together they run technical diplomas, a scientific research camp, olympiad training, and scholarship mentorship.\ngraph TD MF[Mulham Fetna: Renaissance Engineer] style MF fill:#2a004e,color:#fff,stroke:#146c94,stroke-width:4px NB[Neurobotics: Integrated Tech] BL[Boundless: Academic Services] style NB fill:#20002c,color:#fff,stroke:#146c94 style BL fill:#146c94,color:#fff,stroke:#2a004e MF --\u003e NB MF --\u003e BL MF --\u003e VOL[Strategic Volunteering \u0026 Leadership] MF --\u003e EDU[Academic Foundation] NB --\u003e NA[Neurobotics Academy] NB --\u003e NET[Robotics Engineering Team] NB --\u003e ROS[ROS Arabic Community] NA --\u003e P7[Python Mastery: 7+ Cohorts] NA --\u003e MLO[MLOps Mastery: Onsite Aleppo] NA --\u003e HW[HerWill USA: Curriculum Design] P7 --- PA[Paper Airplanes: Mentor of Semester] BL --\u003e SRC[Scientific Research Camp] BL --\u003e BEC[Boundless Elite Club] BL --\u003e SM[Scholarship Mentorship] SRC --- SND[Sanad / UNFPA Partnership] BEC --- AMB[ICSC \u0026 IAAC Ambassador] SM --\u003e UWC[3x UWC Final Acceptances] VOL --\u003e SPN[SyrProNet: Mechatronics Advisor] VOL --\u003e DSFG[DSFG: Executive Coordinator] EDU --\u003e AL[Aleppo University: Mechatronics] AL --- TOP[Top of Class: 89%] classDef project fill:#e1f5fe,stroke:#01579b,color:#000; class NA,SRC,BEC,SM,NET project; Education # B.Sc. Mechatronics Engineering — University of Aleppo, 2023–2027 Fourth year. Top of class, 89% cumulative (first two years). Coursework spans mechanical design, embedded systems, control theory, robotics, and industrial automation.\nSyrian Baccalaureate — 2313 / 2400 (96.375%)\nWhere to next # Open Source — merged contributions to OpenCV and OpenDR. Learn — courses, workshops, and free roadmaps. Ventures — Neurobotics and Boundless. Work With Me — engineering, consulting, and mentorship. Timeline — the journey, year by year. Full experience, affiliations, and certifications: see my CV.\nCommon questions # What is Mulham Fetna\u0026#39;s background? He is a fourth-year mechatronics engineering student at the University of Aleppo, top of his class at 89%, and simultaneously a practising engineer, researcher, and trainer. He founded two Syrian ventures — Neurobotics, for engineering and technical education, and Boundless, for academic services — and has been teaching since 2022. What languages does Mulham Fetna speak? Arabic is his native language and he is fluent in English. He teaches and delivers workshops in both, and founded the Arabic-language ROS developer community. Why does Mulham Fetna run his own website instead of LinkedIn? This site is his platform-independent professional record, built after losing access to LinkedIn. It is designed to outlast any single platform and to remain under his own control. What has Mulham Fetna achieved as a trainer? He has taught 500+ learners across all programs, 2022–2026, coached students to 65 olympiad medals in 2024 as one of 20 nationally selected trainers, and co-led a 40-hour scientific research camp partnered with SANAD and UNFPA Syria. Get in touch: contact@mulhamfetna.com\n","externalUrl":null,"permalink":"/about/","section":"About Mulham Fetna","summary":"Mulham Fetna # Mechatronics \u0026 AI/ML Engineer · Researcher · Technical Trainer\nI build robotics and machine-learning systems, publish research, and train the next generation of Syrian engineers and researchers — from Aleppo.\nMechatronics engineering student at the University of Aleppo (fourth year, top of class at 89%). Open-source contributor with 3 pull requests merged into OpenCV and OpenDR — two of them into OpenCV core.\nTeaching since 2022 across Syria, Lebanon, Palestine, and internationally through Paper Airplanes and HerWill — 500+ learners across all programs, 2022–2026.\nThis site is my platform-independent professional record — built after losing access to LinkedIn, and designed to outlast any single platform.\nThe mission # I work at the intersection of things that are usually kept apart: mechatronics, data science, MLOps, research, STEM training, and open source.\nSyria’s engineering talent is not the bottleneck — access is. Access to a research culture, to global open-source communities, to the tools and the credentials that turn ability into opportunity. Everything I build is aimed at that gap, and at making the path I had to improvise into one that others can simply follow.\n","title":"About Mulham Fetna","type":"about"},{"content":" mulhamfetna/ai-cup-2026-bird-classification Python 0 0 🦅 AI Cup 2026 Bird Classification # Physics-informed machine learning for wind turbine bird strike prevention\nIndustrial AI system combining radar trajectory analysis with gradient boosting ensembles for real-time bird species classification. Features 59+ kinematic features, gliding ratio analysis, and turbine safety controller logic.\n🎯 System Architecture # graph LR A[Radar Sensor] --\u0026gt; B[WKB Parser] B --\u0026gt; C[Feature Engine] C --\u0026gt; D[Kinematic Features] C --\u0026gt; E[Radar Features] C --\u0026gt; F[Temporal Features] D --\u0026gt; G[ML Ensemble] E --\u0026gt; G F --\u0026gt; G G --\u0026gt; H[Bird Classifier] H --\u0026gt; I[Safety Controller] I --\u0026gt; J{Threat Level} J --\u0026gt;|Critical| K[Turbine Shutdown] J --\u0026gt;|Elevated| L[Reduce Speed] J --\u0026gt;|Low| M[Normal Operation] 📊 Problem Description # Multi-class classification of bird radar tracks into 9 categories:\nClutter - Non-bird radar returns Cormorants - Large water birds Pigeons - Urban/common birds Ducks - Waterfowl Geese - Large migratory birds Gulls - Coastal/seabirds (58% of dataset) Birds of Prey - Hawks, eagles (high-altitude gliders) Waders - Shore birds Songbirds - Small passerines (19% of dataset) 🚀 Quick Start # Setup Environment # # Activate virtual environment source .venv/bin/activate # Linux/Mac # or .venv\\Scripts\\activate # Windows # Install dependencies (already done) pip install -r requirements.txt Run Baseline Model # # Fast baseline model (~ 2 minutes) python run_baseline.py Output: outputs/baseline_submission.csv Performance: Average OOF Log Loss: 0.1495\nRun Advanced Ensemble (Optional) # # Full pipeline with advanced feature engineering (~ 15-30 minutes) python run_pipeline.py Output: outputs/submission.csv\nTest Safety Controller # # Demo industrial safety logic python src/safety_controller.py Output: Threat assessment reports for various bird detection scenarios\n📁 Project Structure # . ├── data/ # Competition data │ ├── train.csv # Training data (2,601 tracks) │ ├── test.csv # Test data (1,872 tracks) │ └── sample_submission.csv # Submission format ├── src/ # Source code modules │ ├── features.py # Feature engineering (kinematic, trajectory, radar) │ ├── train.py # Ensemble training (XGBoost, LightGBM, CatBoost) │ └── safety_controller.py # Industrial turbine safety logic ├── notebooks/ # Analysis notebooks │ └── 01_trajectory_visualization.ipynb ├── outputs/ # Model outputs and submissions ├── models/ # Saved model checkpoints ├── run_baseline.py # Quick baseline model script ├── run_pipeline.py # Full ensemble pipeline └── requirements.txt # Python dependencies 🔧 Features # Baseline Model (14 features) # Radar signatures: bird size, airspeed, altitude (min/max/range) Temporal: duration, hour of day, day of week, month, cyclical encoding Trajectory: approximate length from WKB hex string Observer: number of birds observed Advanced Model (59+ features) # Trajectory parsing: WKB geometry decoding for spatial coordinates Kinematic features (22 physics-based): Velocity: 3D velocity, horizontal/vertical components, variance, coefficient of variation Acceleration: Mean, max, variance (flight pattern indicator) Flight patterns: Flapping vs gliding detection Climb dynamics: Climb rate, descent rate Gliding ratio: L/D ratio (horizontal distance / vertical drop) - distinguishes Birds of Prey Altitude efficiency: Distance traveled per meter of altitude change Path geometry: Turn radius, curvature, tortuosity RCS proxy features: Time interval variance, sampling frequency, track density Spatial statistics: Distance, straightness, angular changes Temporal patterns: Cyclical time encoding Safety Controller Module # Industrial logic layer for wind turbine bird strike prevention:\nThreat evaluation: Real-time risk assessment by species and proximity Species risk profiles: High/medium/low impact classification Control actions: Turbine shutdown, speed reduction, or normal operation Batch processing: Multi-detection threat prioritization 🎯 Models \u0026amp; Performance # Baseline # Algorithm: XGBoost Features: 14 basic features Training: 3-fold cross-validation Speed: ~2 minutes Score: 0.1495 OOF Log Loss Per-class: Best - Clutter (0.04), Ducks (0.06) | Hardest - Gulls (0.43), Songbirds (0.30) Advanced Ensemble # Algorithms: XGBoost + LightGBM + CatBoost Features: 59+ physics-based features (22 kinematic + RCS proxies) Training: 5-fold cross-validation per model Ensemble: Average predictions across all models and folds Speed: ~15-30 minutes Key differentiators: Gliding ratio, velocity variance, RCS proxies 🐳 Docker Deployment # Production-ready containerization with multi-stage builds:\n# Build and run baseline docker build --target production -t bird-classifier:latest . docker run -v $(pwd)/data:/app/data bird-classifier:latest # Development with Jupyter docker-compose up jupyter # Access at http://localhost:8888 # Full training pipeline docker-compose up trainer See DOCKER.md for complete deployment guide.\nExpected improvement: Lower log loss through ensemble diversity 📈 Results # Per-Class Performance (Baseline) # Class OOF Log Loss Clutter 0.0398 Cormorants 0.0713 Pigeons 0.0802 Ducks 0.0629 Geese 0.1086 Gulls 0.4336 Birds of Prey 0.1235 Waders 0.1228 Songbirds 0.3030 Average 0.1495 Insights:\nClutter is easiest to predict (low log loss) Gulls and Songbirds are most challenging (high log loss) - likely due to: Large class imbalance (Gulls: 58% of training data) High intra-class variability in flight patterns 📝 Usage Examples # Generate Submission # # Quick baseline python run_baseline.py # → outputs/baseline_submission.csv # Advanced ensemble python run_pipeline.py # → outputs/submission.csv Load and Analyze Results # import pandas as pd # Load submission submission = pd.read_csv(\u0026#39;outputs/baseline_submission.csv\u0026#39;) # Check prediction probabilities print(submission.head()) # Verify probabilities sum (should be close to 1 for each row) print(submission.iloc[:, 1:].sum(axis=1).describe()) 🔍 Data Insights # Training samples: 2,601 radar tracks\nTest samples: 1,872 radar tracks\nClass distribution: Highly imbalanced\nGulls: 58% (1,503 samples) Songbirds: 19% (483 samples) Pigeons, Waders, Birds of Prey: 3-5% each Clutter, Geese, Ducks, Cormorants: \u0026lt;3% each Feature types:\nWKB-encoded spatial trajectories Radar measurements (size, speed, altitude) Temporal metadata (timestamps) Observer annotations (train only) 🚧 Future Improvements # Class imbalance handling:\nSMOTE or class weights Stratified sampling Feature engineering:\nFourier transforms of trajectories Statistical moments of movement patterns Weather/environmental features if available Model enhancements:\nNeural networks for trajectory sequences Attention mechanisms for temporal patterns Stacking with meta-learners Ensemble optimization:\nWeighted averaging based on validation performance Stacking ensemble (train meta-model on OOF predictions) 📄 License # Competition project for educational purposes.\nmulhamfetna/ai-cup-2026-bird-classification Python 0 0 ","externalUrl":null,"permalink":"/projects/ai-cup-2026-bird-classification/","section":"Projects","summary":" mulhamfetna/ai-cup-2026-bird-classification ","title":"AI Cup 2026 Bird Classification","type":"projects"},{"content":"","externalUrl":null,"permalink":"/authors/","section":"Authors","summary":"","title":"Authors","type":"authors"},{"content":" Career Journey # A visual timeline of my professional and academic path from 2020 to present.\n2026 (Current) # January - Present # Open Source: two fixes merged into OpenCV core (Unicode and UTF-8 path handling) Open Source: 3 merged PRs into OpenCV and OpenDR (from 30\u0026#43; submitted) Neurobotics: 4th cohort of Python Mastery in progress 2025 # January # Founded Neurobotics — Engineering services \u0026amp; education company Founded Boundless — Academic services company Awards # Top 3 — Ramadan Initiatives Award 2026 — Directorate of Development Academic # Research Mentorship Program RP1 — Accepted participant 2024 # Teaching # Paper Airplanes: Volunteer of the Month (Fall 2025) Syrian Scientific Olympiad: 65 Olympic medals as coach (12 gold, 21 silver, 32 bronze) Neurobotics Academy: Delivered 7+ Python and MLOps cohorts Projects # Sentiment Analysis Dashboard (Local + Cloud hybrid) Real-Time Edge AI Emotion Detection Local LLM Deployment with RAG N8N Automation Server 2023 # Education # Started Bachelor of Mechatronics Engineering, Aleppo University (4th year of 5) Teaching # Paper Airplanes — Python \u0026amp; MLOps instructor (Women in Tech Program) Syrian National Olympiad — Physics \u0026amp; Robotics coach HerWill — Data Engineering curriculum designer Achievements # Top of class (89% cumulative grade) 2022 # Work # Ala\u0026rsquo;a Screens Company: Embedded Systems Designer (June - October) P10 DMD screen applications Basketball stadium systems School bus displays Education # Syrian Baccalaureate — 2313/2400 (96.375%) 2020-2021 # Freelance # Engineering Consultant: 20+ projects 3 graduation projects 1 master\u0026rsquo;s thesis Full development cycle (design to prototyping) Education Timeline # Year Degree Institution Achievement 2027 Bachelor of Mechatronics Aleppo University Expected graduation 2023-2027 Mechatronics Engineering Aleppo University Top of class (89%) 2022 Syrian Baccalaureate High School 96.375% (2313/2400) Skills Progression # gantt title Skills Development Timeline dateFormat YYYY-MM section Programming Python Basics :2020-01, 2021-06 C++ Fundamentals :2021-01, 2022-06 Advanced Python :2022-01, 2024-12 section AI/ML Data Analysis :2022-01, 2023-06 Machine Learning :2023-01, 2024-06 Edge AI \u0026amp; RAG :2024-01, 2026-01 section Robotics Arduino Basics :2020-01, 2021-06 ROS Fundamentals :2021-06, 2023-12 Control Systems :2023-01, 2026-01 Key Milestones # Year Milestone 2020 Started freelancing 2022 First professional job 2023 Started university 2024 First open source PR 2025 Founded two companies 2026 Fixes merged into OpenCV core ","externalUrl":null,"permalink":"/about/timeline/","section":"About Mulham Fetna","summary":"Career Journey # A visual timeline of my professional and academic path from 2020 to present.\n2026 (Current) # January - Present # Open Source: two fixes merged into OpenCV core (Unicode and UTF-8 path handling) Open Source: 3 merged PRs into OpenCV and OpenDR (from 30+ submitted) Neurobotics: 4th cohort of Python Mastery in progress 2025 # January # Founded Neurobotics — Engineering services \u0026 education company Founded Boundless — Academic services company Awards # Top 3 — Ramadan Initiatives Award 2026 — Directorate of Development Academic # Research Mentorship Program RP1 — Accepted participant 2024 # Teaching # Paper Airplanes: Volunteer of the Month (Fall 2025) Syrian Scientific Olympiad: 65 Olympic medals as coach (12 gold, 21 silver, 32 bronze) Neurobotics Academy: Delivered 7+ Python and MLOps cohorts Projects # Sentiment Analysis Dashboard (Local + Cloud hybrid) Real-Time Edge AI Emotion Detection Local LLM Deployment with RAG N8N Automation Server 2023 # Education # Started Bachelor of Mechatronics Engineering, Aleppo University (4th year of 5) Teaching # Paper Airplanes — Python \u0026 MLOps instructor (Women in Tech Program) Syrian National Olympiad — Physics \u0026 Robotics coach HerWill — Data Engineering curriculum designer Achievements # Top of class (89% cumulative grade) 2022 # Work # Ala’a Screens Company: Embedded Systems Designer (June - October) P10 DMD screen applications Basketball stadium systems School bus displays Education # Syrian Baccalaureate — 2313/2400 (96.375%) 2020-2021 # Freelance # Engineering Consultant: 20+ projects 3 graduation projects 1 master’s thesis Full development cycle (design to prototyping) Education Timeline # Year Degree Institution Achievement 2027 Bachelor of Mechatronics Aleppo University Expected graduation 2023-2027 Mechatronics Engineering Aleppo University Top of class (89%) 2022 Syrian Baccalaureate High School 96.375% (2313/2400) Skills Progression # gantt title Skills Development Timeline dateFormat YYYY-MM section Programming Python Basics :2020-01, 2021-06 C++ Fundamentals :2021-01, 2022-06 Advanced Python :2022-01, 2024-12 section AI/ML Data Analysis :2022-01, 2023-06 Machine Learning :2023-01, 2024-06 Edge AI \u0026 RAG :2024-01, 2026-01 section Robotics Arduino Basics :2020-01, 2021-06 ROS Fundamentals :2021-06, 2023-12 Control Systems :2023-01, 2026-01 Key Milestones # Year Milestone 2020 Started freelancing 2022 First professional job 2023 Started university 2024 First open source PR 2025 Founded two companies 2026 Fixes merged into OpenCV core ","title":"Career Journey","type":"about"},{"content":"Mechatronics and AI/ML engineer, researcher, and technical trainer — Aleppo, Syria.\n4 years teaching · 500\u0026#43; learners across all programs, 2022–2026 · 3 merged PRs into OpenCV and OpenDR · 2 companies founded\nContact: contact@mulhamfetna.com · ORCID 0009-0006-4432-798X\nProfessional Experience # Data Science Researcher | BeInMedia (Kuwait) | 2026 – Present # Full-time data science research for a Kuwait-based media company, focused on qualitative analysis — turning unstructured media and audience data into evidence that a business can act on.\nFounder \u0026amp; CEO | Neurobotics — Integrated Technology Solutions | Jan 2025 – Ongoing # Vision: Engineering minds and machines. Bridging academic theory and industrial mechatronics practice.\nDivisions\nNeurobotics Academy — Python, Data Engineering, and MLOps diplomas. 2 paid cohorts delivered, 20+ students across Syria. Neurobotics Engineering Team — custom mechatronics and robotics solutions. Neurobotics Junior — STEM labs for KG1 through high school. ROS Arabic Community — founded and manage the Arabic-language community for ROS developers. Key achievements\nDelivered 7+ Python and MLOps cohorts with documented student outcomes and published testimonials. Developed the Personal Branding 101 and STEM Junior curriculum frameworks from scratch. Top 3 Ramadan Initiative Award 2026 — Directorate of Development, for excellence in educational services. Invited to deliver a hybrid Professional Personal Branding programme at Homs University IT Department. Founder \u0026amp; CEO | Boundless — Academic Services | Jan 2025 – Ongoing # Vision: Removing barriers to global academic excellence for Syrian and Arab students.\nDivisions\nScientific Research Camp (SRC1) — 75 enrolled, 35\u0026#43; graduated in the first cohort, across engineering, medical, and agriculture research tracks. One publication under review. Boundless Elite Club — international olympiads (IAAC, IYMC, IOI). IPHO national team trainer; ICSC/IAAC ambassador. Scholarship Mentorship — 3 confirmed UWC final acceptances; students placed toward Chevening, DAAD, and GKS. Academic Mentorship — standardized exams (SAT, IELTS, TOEFL, GRE) and professional personal branding. Key achievements\nPartnered with SANAD and UNFPA Syria for the Scientific Research Camp. Ambassador for Aspire Leaders x Harvard and McKinsey Forward. Built the first structured Arabic-language research methodology guide for undergraduates in Aleppo. Python, Data Engineering \u0026amp; MLOps Instructor — Neurobotics Academy # Designed and delivered the complete Python Zero-to-Data-Engineering curriculum across 3 paid cohorts: core Python, data engineering, APIs, async systems, and production portfolio delivery — 36 hours over 3 months.\nPython \u0026amp; MLOps Instructor — Paper Airplanes, Women in Tech # Delivered Python Level 1, Python Level 2, and MLOps full-stack AI to 60\u0026#43; women across 6 countries — the Arab world, Turkey, and Central Asia.\nVolunteer of the Month — Fall 2025: Facebook · Instagram · LinkedIn Data Engineering Curriculum Designer — HerWill # Designed and delivered a full Python data-engineering curriculum under a volunteering contract with HerWill, a US-based women-in-STEM organization.\nSTEM Instructor for GCC Students — Modarby # Delivered STEM sessions to GCC-based students through the Modarby platform.\nSTEM \u0026amp; Programming Trainer — Syrian Virtual University / Distinction and Creativity Agency # One of 20 selected national trainers for the Kids and Adolescents Programming Marathon (KPM), delivering programming fundamentals to 350+ students across Syria. Coached students to 65 olympiad medals (12 gold, 21 silver, 32 bronze) in 2024.\nTechnical Expertise \u0026amp; Research # Open Source Contributions # 3 merged pull requests into OpenCV and OpenDR, from 30\u0026#43; submitted. Full detail: Open Source.\nMerged contributions — each links to the accepted pull request:\nPull request Project What it fixed opencv#28935 OpenCV (core) Unicode temp-path handling on Windows opencv#28880 OpenCV (highgui/Qt) UTF-8 window names preserved in fallback paths opendr#522 OpenDR Python 3.12 compatibility Two of these landed in OpenCV core — one of the most widely used computer-vision libraries in the world. I have also submitted PRs to opencv-python and Ultralytics that were not accepted.\nGitHub: github.com/mulhamfetna\nTechnical Stack # Languages: Python (advanced), C++ (advanced), C, Shell AI/ML: PyTorch, TensorFlow, OpenCV, YOLO, Ultralytics Robotics: ROS/ROS2, Gazebo, control systems (MPC, PID, SO(3)) Embedded: Arduino, STM32, ESP32, PCB design (KiCad, EasyEDA) DevOps: Docker, N8N, FastAPI, Linux\nFull matrix: Skills.\nEducation # B.Sc. Mechatronics Engineering — University of Aleppo | 2023–2027 # Currently in fourth year. Top of class — 89% cumulative grade (first two years). Coursework: mechanical design, embedded systems, control theory, robotics, industrial automation. Syrian Baccalaureate | 2022–2023 # Score: 2313 / 2400 (96.375%) Affiliations # Mechatronics Advisor — Syrian Professional Network (SyrProNet) Technical advisor on mechatronics within Syria\u0026rsquo;s largest professional technical network.\nExecutive Coordinator — Northern Syria Team, DSFG Coordinating technical and development initiatives across northern Syria through the German–Syrian Research Foundation (Deutsch-Syrische Forschungsgesellschaft e.V.).\nMember — Syrian Entrepreneurs Union Technical and educational entrepreneurship track.\nMember — SYNC / Digital Syria Digital transformation and technology education initiatives.\nVolunteering # Volunteer Instructor — Paper Airplanes, Women in Tech | 2024–Ongoing Python and MLOps courses to women in conflict-affected regions. Volunteer of the Month, Fall 2025.\nVolunteer Curriculum Designer — HerWill | 2024 Designed a full Python data-engineering curriculum on a voluntary basis.\nWUDC Debater — British Parliamentary format | 2023–Ongoing Competitive debater in the World Universities Debating Championship format.\nNational Programming Trainer — KPM / SVU / DCA | 2022–2023 One of 20 nationally selected trainers delivering programming education to children and adolescents across Syria, under the Kids Programming Marathon — organized jointly by the Syrian Virtual University, the Distinction and Creativity Agency, and the Syrian Science Olympiad.\nLicenses, Certifications, Honors \u0026amp; Awards # Top 3 — Ramadan Initiatives Award 2026 Directorate of Development, for excellence in educational services through Neurobotics Academy and Boundless Elite Club.\nCS50 Certified Instructor — Harvard University Authorized to deliver Harvard\u0026rsquo;s CS50 curriculum independently.\nProject Management Professional (PMP) Preparation Completed structured PMP preparation coursework.\nMonitoring, Evaluation, Accountability and Learning (MEAL) Structured preparation in MEAL methodology for research and programme management.\nData Collection — Certified Structured data-collection methodology for applied research contexts.\nMcKinsey Forward Program Completed McKinsey\u0026rsquo;s professional development programme for emerging leaders.\neCornell — Take the Lead Program Leadership development, Cornell University\u0026rsquo;s online division.\nAspire Leaders Program — Harvard University Selected participant; currently serving as Aspire Leaders Ambassador.\nResearch Mentorship Program — RP1 | 2026 Accepted participant, focused on academic research methodology and the publication track.\nOrganizations # Syrian Professional Network (SyrProNet) — Mechatronics Advisor DSFG — Executive Coordinator, Northern Syria Team Paper Airplanes — Volunteer Instructor, Women in Tech Program HerWill — Volunteer Curriculum Designer Syrian Entrepreneurs Union — Member SYNC / Digital Syria — Member Aspire Leaders — Ambassador McKinsey Forward — Alumnus Research Mentorship Program RP1 — Participant, 2026 ROS Arabic Community — Founder Boundless — Founder \u0026amp; CEO Neurobotics — Founder \u0026amp; CEO Courseing Platform — Partner Instructor ","externalUrl":null,"permalink":"/work-with-me/cv/","section":"Work With Me","summary":"Mechatronics and AI/ML engineer, researcher, and technical trainer — Aleppo, Syria.\n4 years teaching · 500+ learners across all programs, 2022–2026 · 3 merged PRs into OpenCV and OpenDR · 2 companies founded\nContact: contact@mulhamfetna.com · ORCID 0009-0006-4432-798X\nProfessional Experience # Data Science Researcher | BeInMedia (Kuwait) | 2026 – Present # Full-time data science research for a Kuwait-based media company, focused on qualitative analysis — turning unstructured media and audience data into evidence that a business can act on.\nFounder \u0026 CEO | Neurobotics — Integrated Technology Solutions | Jan 2025 – Ongoing # Vision: Engineering minds and machines. Bridging academic theory and industrial mechatronics practice.\nDivisions\nNeurobotics Academy — Python, Data Engineering, and MLOps diplomas. 2 paid cohorts delivered, 20+ students across Syria. Neurobotics Engineering Team — custom mechatronics and robotics solutions. Neurobotics Junior — STEM labs for KG1 through high school. ROS Arabic Community — founded and manage the Arabic-language community for ROS developers. Key achievements\nDelivered 7+ Python and MLOps cohorts with documented student outcomes and published testimonials. Developed the Personal Branding 101 and STEM Junior curriculum frameworks from scratch. Top 3 Ramadan Initiative Award 2026 — Directorate of Development, for excellence in educational services. Invited to deliver a hybrid Professional Personal Branding programme at Homs University IT Department. Founder \u0026 CEO | Boundless — Academic Services | Jan 2025 – Ongoing # Vision: Removing barriers to global academic excellence for Syrian and Arab students.\n","title":"Curriculum Vitae","type":"work-with-me"},{"content":"\u0026ldquo;Data\u0026rdquo; and \u0026ldquo;AI\u0026rdquo; get used as if they were one job. They are at least four, they need different people, and confusing them is how organizations waste a year and a budget.\nThis workshop takes the field apart — for people who will use, fund, hire for, and judge data and AI work, not build it themselves.\nFormat # 4 hours, on-site, highly interactive.\nDelivered in 2026 for SyrProNet, the Syrian Professional Network, to an audience of senior professionals — founders, consultants, project managers, academics, and decision-makers.\nWhat we cover # The four roles, honestly distinguished. Data analysis — what happened. Data engineering — the plumbing that makes everything else possible, and the part almost everyone underfunds. Data science — why it happened and what is likely next. Machine learning and AI — systems that decide or generate, including the ones that fail quietly.\nWhat each one actually needs. The skills, tools, timeline, and realistic cost. What you can expect in three months versus three years.\nHow to judge the work. The questions to ask a data hire, a vendor, or a consultant that separate substance from theatre. Why \u0026ldquo;we use AI\u0026rdquo; is not an answer, and what a good answer sounds like.\nWhere AI genuinely helps — and where it does not. Including the uncomfortable cases: when the data does not exist, when the problem is organizational rather than technical, and when a spreadsheet is the correct solution.\nWho it\u0026rsquo;s for # Founders, managers, consultants, academics, and decision-makers who need to be literate, not technical. No coding, no mathematics prerequisite.\nAlso useful for students choosing which of these paths to actually enter.\nBook this workshop # Available for professional networks, companies, and universities. Get in touch or email contact@mulhamfetna.com.\nCommon questions # Do I need a technical background for this workshop? No. It is built for decision-makers — founders, managers, consultants, and academics — who need to understand and judge data and AI work without building it. There is no coding and no mathematics prerequisite. What is the difference between data analysis, data engineering, and data science? Data analysis explains what happened. Data engineering builds the pipelines and infrastructure that make any analysis possible, and is the most commonly underfunded of the three. Data science explains why something happened and what is likely to happen next. Machine learning and AI build systems that decide or generate. These are different jobs needing different people, and treating them as one is a common and expensive mistake. Who delivers this workshop? Mulham Fetna, a mechatronics and AI/ML engineer based in Aleppo, Syria. He delivered it in 2026 as a four-hour on-site session for SyrProNet, the Syrian Professional Network, to senior professionals including founders, consultants, project managers, and academics. ","externalUrl":null,"permalink":"/workshops-camps/data-ai-careers-behind-it/","section":"Workshops \u0026 Camps","summary":"Data Analysis/Engineering/Science | Machine Learning \u0026 AI - under the telescope","title":"Data \u0026 AI - The Careers Behind It","type":"workshops-camps"},{"content":" mulhamfetna/kaggle-playground-series-s6e4 Python 0 0 Kaggle Playground Series S6E4 - From Competition Model to Deployable API # This repository demonstrates a full path from Kaggle tabular modeling to a production-style inference service.\nWhat this project proves # Competitive ML workflow for playground-series-s6e4 (Predicting Irrigation Need). Reproducible training pipeline with configurable compute budgets. Deployable serving layer (FastAPI + Docker), not just notebook experimentation. Repository structure # kaggle_gpu_submission_workflow.ipynb\nKaggle Notebook workflow to train/evaluate and generate submission_ready.csv.\ntrain_advanced_and_submit.py\nAdvanced competition trainer (CatBoost + CV + optional search + report JSON).\nkaggle_submission_guide.md\nStep-by-step notebook submission guide.\ntrain_api_model.py\nTrains a CPU CatBoost model and exports serving artifacts to model_artifacts/.\napp/main.py\nFastAPI service with:\nGET /health GET /schema POST /predict Dockerfile + requirements.txt\nContainerized API deployment stack.\nArchitecture # flowchart LR A[train.csv] --\u0026gt; B[Feature selection + categorical preprocessing] B --\u0026gt; C[CatBoost training] C --\u0026gt; D[Model artifact .cbm] C --\u0026gt; E[Metadata JSON] D --\u0026gt; F[FastAPI inference service] E --\u0026gt; F F --\u0026gt; G[/predict -\u0026gt; Irrigation_Need class/] Kaggle workflow # Open the competition notebook environment. Run kaggle_gpu_submission_workflow.ipynb. Produce: /kaggle/working/submission_ready.csv /kaggle/working/model_report.json Submit in competition UI. API training and local serving # 1) Install dependencies # pip install -r requirements.txt 2) Train and export API artifacts # python train_api_model.py --data-dir . --output-dir model_artifacts --iterations 700 --seed 42 3) Run FastAPI # uvicorn app.main:app --host 0.0.0.0 --port 8000 4) Test endpoints # curl http://127.0.0.1:8000/health curl http://127.0.0.1:8000/schema Docker deployment # Build image (after generating model_artifacts/):\ndocker build -t irrigation-api:latest . Run container:\ndocker run --rm -p 8000:8000 irrigation-api:latest Business framing # This repo is designed as a public proof asset for ML/Data Engineering work:\nmeasurable leaderboard performance reproducible experiments deployable inference interface clear operational documentation mulhamfetna/kaggle-playground-series-s6e4 Python 0 0 ","externalUrl":null,"permalink":"/projects/kaggle-playground-series-s6e4/","section":"Projects","summary":" mulhamfetna/kaggle-playground-series-s6e4 ","title":"Kaggle Playground Series S6E4 - From Competition Model to Deployable API","type":"projects"},{"content":"Practical, project-based technical education — taught in Arabic and English, online and in person.\nI have taught 500\u0026#43; learners across all programs, 2022–2026: university cohorts, women re-skilling into tech across six countries, and nationally selected olympiad students.\nEverything I teach is built on the same principle — you finish with something you built, not just notes you took.\nStart here # Courses — Cohort-based diplomas with graduation projects: Python Mastery, Python to Data Engineering, and MLOps to Full-Stack AI Engineer. Assessed, not just attended.\nWorkshops — Short, intensive programmes: Personal Branding 101, Mechatronics \u0026amp; Robotics 101, and Data \u0026amp; AI Careers.\nRoadmaps — Free, self-paced paths you can follow without me: a 12-level Mechatronics roadmap and a Python-to-MLOps roadmap, both bilingual.\nTutorials — Free written guides from real engineering work.\nWho this is for # Beginners with no programming background, university students who want the practical skills their degree skips, and career switchers moving into AI and data engineering. Courses run in Arabic and English.\nWhat students say # Read student outcomes and testimonials — including results from the Paper Airplanes Women in Tech programme and Neurobotics Academy cohorts.\nReady to start? # See the current courses, or book a mentorship session if you want a path built around you.\nCommon questions # What courses does Mulham Fetna teach? Cohort-based diplomas in Python (from zero to job-ready), Python to Data Engineering, and MLOps to Full-Stack AI Engineer, plus short workshops in personal branding, mechatronics and robotics, and data and AI careers. Free self-paced roadmaps in mechatronics and Python-to-MLOps are also available. Does Mulham Fetna teach in Arabic? Yes. Courses and workshops are delivered in both Arabic and English, online and in person. He also founded the Arabic-language ROS (Robot Operating System) developer community. Who are these courses for? Beginners with no programming background, university students who want the practical skills a degree tends to skip, and career switchers moving into AI and data engineering. No prior experience is required for the entry-level Python track. Are the courses assessed, or just attended? Assessed. Cohorts run on an 80% attendance requirement and finish with a graduation project, so students leave with something they built rather than notes they took. How do I enroll? Browse the current courses and apply through the form on the course page. For one-to-one guidance instead, book a mentorship session, or email contact@mulhamfetna.com. ","externalUrl":null,"permalink":"/learn/","section":"Learn With Me","summary":"Practical, project-based technical education — taught in Arabic and English, online and in person.\nI have taught 500+ learners across all programs, 2022–2026: university cohorts, women re-skilling into tech across six countries, and nationally selected olympiad students.\nEverything I teach is built on the same principle — you finish with something you built, not just notes you took.\nStart here # Courses — Cohort-based diplomas with graduation projects: Python Mastery, Python to Data Engineering, and MLOps to Full-Stack AI Engineer. Assessed, not just attended.\nWorkshops — Short, intensive programmes: Personal Branding 101, Mechatronics \u0026 Robotics 101, and Data \u0026 AI Careers.\nRoadmaps — Free, self-paced paths you can follow without me: a 12-level Mechatronics roadmap and a Python-to-MLOps roadmap, both bilingual.\nTutorials — Free written guides from real engineering work.\nWho this is for # Beginners with no programming background, university students who want the practical skills their degree skips, and career switchers moving into AI and data engineering. Courses run in Arabic and English.\nWhat students say # Read student outcomes and testimonials — including results from the Paper Airplanes Women in Tech programme and Neurobotics Academy cohorts.\n","title":"Learn With Me","type":"learn"},{"content":"Most people meet robotics through films, then meet it again through a university syllabus, and the two have almost nothing in common. Neither tells you what the work actually is, where it happens, or how to get into it.\nThis workshop closes that gap — from the fantasy, to the factory floor, to the research lab.\nWhat we cover # What mechatronics actually is. The honest definition: mechanical engineering, electronics, control, and software, forced to cooperate. Why no one of those disciplines is enough on its own, and why the integration is the hard part.\nWhere robots really work. Manufacturing and industrial automation, agriculture, medical and surgical robotics, logistics and warehousing, inspection and hazardous environments, and defence. What each one demands, and which are actually hiring.\nThe stack, demystified. Sensors and actuators, microcontrollers and embedded systems, control theory, and ROS — how a real robot is assembled from parts you can name, not magic.\nResearch versus industry. What a robotics researcher does all day compared with a robotics engineer, and which one you would actually enjoy.\nHow to enter the field. The realistic path: what to learn, in what order, what to build, and what to ignore. This maps directly onto the free 12-level Mechatronics Roadmap — you leave with a plan, not inspiration.\nWho it\u0026rsquo;s for # Engineering students choosing a specialization, beginners curious whether robotics is for them, and anyone who wants an honest picture before committing years to it.\nNo prerequisites. Delivered in Arabic and English.\nKeep going after the workshop # The workshop is the map. The Mechatronics Roadmap is the territory — 12 levels from analog electronics to advanced robotics and reinforcement learning, free and self-paced.\nBook this workshop # Available for universities, student clubs, and schools. Get in touch or email contact@mulhamfetna.com.\nCommon questions # What is mechatronics engineering? Mechatronics is the integration of mechanical engineering, electronics, control systems, and software into a single working machine. No one of those disciplines is sufficient alone — the difficulty, and the discipline, lies in making them cooperate. It is the engineering behind robots, industrial automation, and most modern smart devices. Do I need prior experience to attend? No. The workshop is designed for engineering students choosing a specialization and for complete beginners deciding whether robotics is for them. There are no prerequisites, and it is delivered in Arabic and English. Which industries actually employ robotics engineers? Manufacturing and industrial automation, agriculture, medical and surgical robotics, logistics and warehousing, inspection in hazardous environments, and defence. The workshop covers what each sector demands and which are genuinely hiring, rather than which are simply talked about. How do I start learning robotics? Follow a structured path rather than collecting tutorials. The free 12-level Mechatronics Roadmap on this site runs from analog and digital electronics through embedded controllers and ROS to advanced robotics and reinforcement learning, and is self-paced in Arabic and English. ","externalUrl":null,"permalink":"/workshops-camps/mechatronics-robotics-101/","section":"Workshops \u0026 Camps","summary":"From Marvel Movies Fantasies to Factory Floors - Top Industries and Research","title":"Mechatronics \u0026 Robotics 101","type":"workshops-camps"},{"content":" Level 1: Analog Electronics # Concepts: Voltage (V) \u0026amp; Current (I), AC vs DC, phase shift, Ohm’s Law Wire selection, current sources, fuses Using lab tools: multimeters, oscilloscopes, signal generators, test boards Simulation tools: Multisim, Circuit Wizard, Proteus Components: resistors, capacitors, voltage dividers, KCL/KVL, inductors, transformers, RLC circuits, relays, contactors, diodes, transistors, zeners, opamps, 555 timer, PWM, filters Sensor \u0026amp; actuator basics Resource: Analog Electronics Video Playlist Level 2: Digital Electronics # Concepts:. Number systems: binary, octal, hex Logic gates: basic (AND, OR, NOT), universal (NAND, NOR, XOR, XNOR) Buffer, schmitt trigger, output types (active high/low, high impedance, open drain) Boolean algebra, Karnaugh maps Encoders, decoders, multiplexers/demultiplexers, BCD Adders, subtractors (half/full) Sequential circuits: flip-flops (SR, JK, T, D), debouncing, async/sync counters Memory types and D/A, A/D converters Resource: Digital Electronics Video Playlist Level 3: C/C++ Fundamentals # Concepts: Programming introduction: languages, compilers, IDEs, debugging, linking, libraries Comments, variables, data types, data structures Operators: arithmetic, logic Control flow: conditionals, arrays, strings, loops, functions Resource: Elzero C++ Study Plan Level 4: Arduino \u0026amp; MCU Basics # Concepts: GPIO, timing, direct register access Basic sensors and actuators: IR, ultrasonic, keypad, brushed/brushless/stepper/servo motors, drivers Displays: LCD, OLED, 7-segment, dot-matrix, shift registers (MAX7219, etc) Communication: I2C, UART, SPI, CAN Wireless: Bluetooth, WiFi, mesh, simple RF Tools: Arduino IDE, serial plotter/viewing tools Level 5: Computer Architecture # Concepts: Memory management, CPU/peripheral basics Microarchitecture: buses, pipelines Resource: Standard text: \u0026ldquo;Computer Architecture\u0026rdquo; by Charles Fox or alternatives Level 6: Advanced Embedded \u0026amp; Controllers Perspective # Concepts: Familiarization with ATmega, STM32, ESP32, RP2040, nRF PCB design and layout basics Reading and interpreting datasheets Practical selection criteria, in-depth register and memory mapping, analog/digital converters, counters Comparative study: ATmega vs STM32 vs ESP32 vs RP2040 architecture Skills: Schematic capture, basic board design Level 7: RTOS, ROS, and Robotics Applications # Concepts: Intro to FreeRTOS, Zephyr, or ChibiOS RTOS tasking, scheduling, IPC, semaphores ROS/micro-ROS overview and building distributed robotic systems Bridges between MCUs and ROS nodes Projects: Real-time sensor fusion, multitasking robot, basic mobile robot with ROS navigation stack Level 8: Python \u0026amp; Edge AI Overview # Concepts: Python basics for embedded/robotics engineers Edge AI topics: computer vision, NLP, TTS, STT Frameworks: TensorFlow Lite, ONNX Runtime, basic model deployment on Pi or MCU Level 9: Control Theory # Concepts: Classic control: PID, lead/lag, frequency response Advanced control: State-space, LQR, optimal/adaptive control Level 10: Robotics Math # Concepts: Inverse kinematics and forward kinematics Transformation matrices, Denavit–Hartenberg (D-H) notation Level 11: Deep ROS Applications # Concepts: Advanced ROS: MoveIt (manipulators), SLAM (mapping), Nav2, tf, rviz visualization Level 12: Advanced Robotics \u0026amp; Reinforcement Learning # Topics: Reinforcement learning: Isaac Lab, Isaac Sim Simulation and high-level robot autonomy Parallel Skills Development # At all levels: KiCad (PCB design), MATLAB, SolidWorks (mechanical CAD) Additional CAM/CAD tools as needed for fabrication or digital twins Industry Pathways (Where to Apply This Roadmap) # Industrial sector: smart assembly and welding arms Medical sector: nanorobots Smart materials: soft robotics Medical sector: surgical robots and robotic arms Prosthetics: artificial limbs Humanoids: cinema and military applications Drones and legged robots (spider-like): services and military Automotive: autonomous vehicles and EV powertrains Aerospace: UAV controls and satellite mechanisms Agriculture: precision farming robots Logistics/warehousing: AGVs and sorting systems Energy: smart grids and wind-turbine automation Consumer electronics: wearables and smart home devices Trend Topics to Track While Learning # Edge AI for embedded systems: NLP, LLMs, computer vision, YOLO, object detection, sentiment detection Reinforcement learning: Isaac ecosystem Mesh networks for IoT devices ROS (Robot Operating System) RTOS (Real-Time Operating Systems) Sensor fusion for navigation systems Digital twins and simulation: Unity/Unreal Engine, NVIDIA Omniverse, MATLAB/Simulink Cybersecurity for robotics/IoT: secure boot, intrusion detection Sustainable mechatronics: bio-inspired design, energy harvesting, recyclable actuators Haptics and teleoperation: force feedback for training and remote operation Multi-robot coordination: SLAM and swarm algorithms content will be updated soon !\nEstimated Timeline (Guided Pace) # Level Focus Estimated Duration 1 Analog Electronics 3-4 weeks 2 Digital Electronics 3-4 weeks 3 C/C++ Fundamentals 3-5 weeks 4 Arduino \u0026amp; MCU Basics 4-6 weeks 5 Computer Architecture 2-3 weeks 6 Advanced Embedded \u0026amp; Controllers 5-7 weeks 7 RTOS, ROS, Robotics Applications 6-8 weeks 8 Python \u0026amp; Edge AI Overview 4-6 weeks 9 Control Theory 4-6 weeks 10 Robotics Math 4-6 weeks 11 Deep ROS Applications 6-8 weeks 12 Advanced Robotics \u0026amp; Reinforcement Learning 6-10 weeks Total expected timeline: ~46-73 weeks (depending on pace, background, and project depth).\nLearning Checkpoints # Checkpoint A (After Levels 1-2) # Build and test a mixed analog/digital mini-circuit. Explain signal flow, logic gates, and measurement workflow. Document your work with schematics and debug notes. Checkpoint B (After Levels 3-4) # Program a complete MCU project (sensor + actuator + display). Use modular C/C++ structure (headers, source files, reusable functions). Demonstrate serial diagnostics and protocol communication (I2C/UART/SPI). Checkpoint C (After Levels 5-6) # Select a microcontroller family for a real use case and justify the choice. Read and apply datasheet sections (timers, interrupts, ADC/PWM). Produce one simple PCB or equivalent hardware design package. Checkpoint D (After Levels 7-8) # Build a ROS-integrated embedded prototype. Run at least one edge AI inference workflow (CV/NLP) on constrained hardware. Document latency, memory, and power tradeoffs. Checkpoint E (After Levels 9-10) # Tune and validate at least one control loop (PID or state-space). Solve one practical kinematics problem with reproducible calculations. Connect control + kinematics to one robotic task scenario. Checkpoint F (After Levels 11-12) # Deliver a full robotics capstone (simulation + hardware/software architecture). Include deployment notes, test logs, and performance evaluation. Publish project documentation suitable for portfolio/interviews. Progress Tracker (Quick Self-Evaluation) # Stage Coverage Target Output A Levels 1-2 Circuit fundamentals demo + measurement notes B Levels 3-4 Embedded mini-system with sensors/actuators C Levels 5-6 Datasheet-driven controller selection + PCB artifact D Levels 7-8 ROS-integrated prototype + edge AI experiment E Levels 9-10 Control + kinematics validated on one robotics task F Levels 11-12 End-to-end capstone with portfolio-grade documentation Downloadables and Paths (Coming Soon) # Mechatronics roadmap printable PDF (single-page visual map). Weekly tracker template (Notion/Sheet format). Lab checklist pack (electronics + embedded debugging). ROS project starter kit template. Capstone documentation template (report + architecture diagram). Level-by-level study path sheet (beginner/intermediate/advanced tracks). Project milestone checklist (submission-ready format). Placeholder links (to be activated soon):\nDownload center: /downloads/ (coming soon) Mechatronics roadmap pack: /downloads/mechatronics-roadmap-pack/ (coming soon) Project templates: /downloads/project-templates/ (coming soon) FAQ # 1. Is this roadmap beginner-friendly? # Yes, but Level 1 assumes you are ready for disciplined technical study and hands-on practice.\n2. Can I skip C/C++ and start with Python only? # Not recommended for embedded and robotics depth. Python helps in higher levels, but C/C++ remains foundational for MCU control.\n3. When should I start ROS? # Start ROS after you are comfortable with embedded basics and real sensor/actuator work (typically around Level 7).\n4. Do I need hardware from day one? # You can begin with simulation, but real hardware should start no later than Level 4 for practical competence.\n5. How much math is required? # You need practical algebra, trigonometry, and linear algebra before deep control/kinematics stages.\n6. Which projects are best for portfolio value? # Projects that combine hardware + firmware + control + documentation + measured performance.\n7. Should I specialize in embedded, ROS, or AI first? # Build a base through Level 8, then specialize based on target role and available project opportunities.\n8. Can this roadmap support job applications? # Yes. If you complete checkpoints and publish 2-4 quality projects with clear engineering reports, it becomes strong interview evidence.\nCall to Action: Consultation and Collaboration # If you want personalized help implementing this roadmap, you can:\nBook a roadmap consultation session (booking link coming soon). Invite me to collaborate on your project through the contact form (form link coming soon). Request technical review for your capstone architecture and execution plan. For upcoming workshops and announcements, follow the main workshops page:\nWorkshops \u0026amp; Camps ","externalUrl":null,"permalink":"/roadmaps/mechatronics-roadmap/","section":"Your Zero-to-Hero Path to Industry-Ready Expertise","summary":"Analog \u0026 Digital Electronics -\u003e … -\u003e  Advanced Embedded \u0026 Controllers Perspective … -\u003e Deep ROS Applications -\u003e  Advanced Robotics \u0026 Reinforcement Learning","title":"Mechatronics Engineering Roadmap","type":"roadmaps"},{"content":" Course Foundation # Course Duration and Scope:\nThis course is designed to be delivered over 27 sessions, each lasting approximately 1 to 1.5 hours. It offers a comprehensive, step-by-step journey from the fundamentals of AI chatbots to advanced machine learning and deep learning techniques. The curriculum spans five levels covering practical use of popular AI chatbots and APIs, local AI model deployment with secure networking, sophisticated Arabic NLP combined with web scraping, multimodal image processing and emotion recognition, and foundational deep learning architectures and algorithms. Each session integrates theory with practical examples and assignments, culminating in projects that synthesize learned skills for real-world AI applications.\nRelated pages to execute this track # Python Data Engineering \u0026amp; MLOps roadmap for full phase-by-phase progression. Courses hub for complementary tracks. Workshops \u0026amp; Camps for practical live sessions. Mentorship Services for one-to-one implementation and project strategy. License and Disclaimer # This course material is designed and provided solely for use within Neurobotics Academy.\nAll rights are reserved by Neurobotics Academy and the course developer Eng. Mulham Fetna, CEO \u0026amp; Founder of Neurobotics Academy.\nCourse Design and Maintenance:\nThis course has been designed, developed, and maintained by Eng. Mulham Fetna and Neurobotics Academy as part of their STEM curriculum development initiatives.\nDate of Development:\nDecember 2025\nRights Reserved:\nAll intellectual property rights, including copyrights, are exclusively reserved for Neurobotics Academy and Eng. Mulham Fetna. Unauthorized use, reproduction, or distribution outside of Neurobotics Academy is strictly prohibited.\nWeeks 1-2: Getting Started with AI Chatbots and APIs # Introduction to AI chatbots, their types, and applications Hands-on demos with ChatGPT, Google Gemini, Claude, Copilot Understanding API concepts and authentication Building simple chatbot apps using paid and free APIs (OpenAI, Grok, Hugging Face) Assignment: Build and compare chatbots with free and paid APIs Weeks 3-4: Local AI Models and Secure Networking # Installing and running local AI models (LLaMA, Ollama, OpenWebUI) Exposing local models via REST APIs with FastAPI Networking basics: localhost, IP addresses, ports Secure remote access using tunneling tools (Twingate, Cloudflare Tunnel) Security best practices: zero-trust models, API keys, rate limiting Weeks 5-6: Arabic Natural Language Processing (NLP) and Web Scraping # Arabic NLP foundations: AraBERT, CAMeL Tools, PyArabic Tokenization, Named Entity Recognition (NER), and sentiment analysis for Arabic text Web scraping tools and techniques (BeautifulSoup, Selenium) for Arabic content collection Integrating scraped data into Arabic NLP pipelines Assignment: Build an Arabic social media sentiment analyzer Weeks 7-8: Image Processing and Multimodal Emotion Detection # Image manipulation basics with OpenCV: resizing, filtering, color spaces Object detection using pretrained YOLO models on images and videos Facial emotion recognition using DeepFace and FER models Combining text and image sentiment for multimodal analysis Assignment: Develop multimodal social media post analysis system Weeks 9-12: Deep Learning Frameworks and Advanced Machine Learning # Neural network fundamentals: perceptrons, activations, forward and backward passes Loss functions and backpropagation explained with examples Supervised learning: regression, classification, decision trees Unsupervised learning: clustering (K-Means, DBSCAN), dimensionality reduction (PCA, t-SNE) Deep architectures: CNNs, RNNs, LSTMs, transfer learning Comparative analysis of algorithms: strengths, weaknesses, and best use cases Final project review and integration across domains Additional Recommendations # Hands-on Practice: Encourage building projects incrementally to reinforce concepts Community Engagement: Join AI and ML communities for peer support and collaboration Version Control: Introduction to Git and GitHub for code management and sharing Ethics and Responsible AI: Discuss implications and safe AI deployment practices Summary Table # Weeks Theme/Topics 1-2 AI Chatbots, APIs, simple chatbot development 3-4 Local AI models, API deployment, networking, security 5-6 Arabic NLP foundations, web scraping, NLP pipeline integration 7-8 Image processing, object detection, emotion detection, multimodal 9-12 Deep learning fundamentals, supervised \u0026amp; unsupervised learning Algorithm pros/cons, project integration, review This course is crafted and delivered by Eng. Mulham Fetna, CEO \u0026amp; Founder of Neurobotics Academy, dedicated to fostering cutting-edge AI \u0026amp; ML expertise for students and professionals.\nCourse Objectives # Course Structure Overview # Your \u0026ldquo;AI/ML from Zero to Hero\u0026rdquo; course spans five levels, progressing from accessible cloud AI to advanced custom models, with 20-30 sessions of 1-1.5 hours each. Allocate roughly 4-6 sessions per level to fit the range, ensuring in-depth coverage: partial explanations with code snippets, integrated real-world examples (e.g., building a customer support bot), and pre-session assignments with hints. End each level with a review session, project assignment (e.g., deploy a multi-API chatbot for Level 1), and culminate in a graduation project like an end-to-end Arabic sentiment analysis app with web scraping and image emotion detection.\nLevel 1: Famous AI Chatbots and APIs # Start with user-facing demos of ChatGPT (versatile for writing/coding, GPT-4o model with voice/tools), Gemini (Google-integrated, multimodal with real-time search/video), Copilot (productivity in Microsoft tools), Claude (long-context reasoning), and Meta AI (casual social use). Key differences include ChatGPT\u0026rsquo;s broad features versus Gemini\u0026rsquo;s deep research and Claude\u0026rsquo;s structured outputs. Transition to APIs: Compare paid OpenAI/Gemini APIs for apps like translators, then free options like Grok API (xAI\u0026rsquo;s witty, real-time responses) and Hugging Face (open models like Llama via inference API).\nLevel 2: Local Models and Networking # Cover running Llama models locally via Ollama or LLM Studio for privacy/offline use, with OpenWebUI for a ChatGPT-like interface. Dive into local APIs (e.g., Ollama endpoints), networking basics (expose via localhost tunneling), and tools like Twingate (secure zero-trust access) or Cloudflare Tunnel (free, easy port forwarding) to access your local AI across devices.[8]\nLevels 3-5: Advanced Topics # Level 3 (NLP): Use Arabert/CAMeL Tools/PyArabic for Arabic text analysis (tokenization, NER, sentiment); integrate web scraping with BeautifulSoup/Selenium for real-time data (e.g., analyze social media posts).\nLevel 4 (Images): Pretrained YOLO for object detection, OpenCV for processing, plus sentiment/emotion models (e.g., FER via DeepFace) on faces/text overlays.\nLevel 5 (Frameworks): TensorFlow/PyTorch basics—neural nets (CNNs/RNNs), loss functions (MSE/cross-entropy), regression/classification, supervised (labeled data, high accuracy but data-intensive) vs. unsupervised (clustering, scalable but less precise)—with pros/cons tables per algorithm.\nSession and Progression Tips # Each session: 20-min partial concepts (e.g., API auth code), 30-min integrated example (e.g., RAG chatbot with LangChain), 20-min assignment (build/deploy with hints like \u0026ldquo;use streamlit for UI\u0026rdquo;). Reviews consolidate via quizzes/projects; track via GitHub for continuity with your bulk-processing workflow. This fits 25 sessions (5/level), scalable to 30 with extras on RAG/agents.\nCourse Plan # Level 1: Famous AI Chatbots and APIs (5 Sessions) # Session 1: Introduction to AI Chatbots # Overview of AI chatbots: What and why? Demo popular AI chatbots as users (ChatGPT, Google Gemini, Claude, Copilot) Key differences in design, use cases, features, and architecture concepts Assignment: Compare chatbot responses on a given topic with reasoning hints Session 2: Using Chatbot APIs (OpenAI, Google, Anthropic) # Introduction to API concepts and authentication Making API calls to ChatGPT and similar platforms Simple chatbot app example with API calls (Python + simple UI) Assignment: Build a simple Q\u0026amp;A using ChatGPT API with hints Session 3: Exploring Free AI APIs (Grok, Open-source APIs) # Overview of free AI API platforms (Grok, Hugging Face, Cohere) Demo API calls and capabilities vs paid APIs Integrate a free API into a mini chatbot or text generation app Assignment: Create a chatbot using a free API and document limitations Session 4: Integrating Multiple APIs and Advanced Features # Combining APIs for enhanced features (e.g., RAG - retrieval augmented generation) Handling errors, rate limits, and streaming results Use-case example: Info bot with multiple API sources Assignment: Extend chatbot to query multiple APIs with hints Session 5: Level 1 Review and Project Assignment # Recap key concepts and compare chatbot types and APIs Level project assignment: Build a customer support chatbot combining paid and free APIs Guidance on project milestones and evaluation criteria Level 2: Local Models and Networking (5 Sessions) # Session 6: Introduction to Local AI Models (LLaMA, Ollama, OpenWebUI) # Understanding local models vs cloud models Installing and running LLaMA-based models offline Demo OpenWebUI for chatting locally Assignment: Set up a local model instance with basic interaction Session 7: Working with Local Model APIs # Exposing local models as REST APIs Simple backend server creation for local AI model calls Assignment: Build a small local API wrapper for a LLaMA model Session 8: Networking Basics for Local AI Access # Intro to networking concepts: localhost, ports, IPs Using tunneling tools: Twingate, Cloudflare Tunnel Secure access to local AI services from other devices Assignment: Connect to a local AI model remotely using tunneling Session 9: Security and Best Practices for Local AI Deployment # Security principles: zero trust, encryption, firewall Practical security setup using Twingate, Cloudflare Assignment: Harden local AI API deployment and test secure access Session 10: Level 2 Review and Project Assignment # Review local model setup, APIs, networking Project: Deploy a local chatbot accessible securely from multiple devices Guidelines and expectations Level 3: NLP for Arabic and Web Scraping (5 Sessions) # Session 11: Arabic NLP Foundations (AraBERT, PyArabic, CAMeL Tools) # Arabic script challenges and preprocessing Tokenization, POS tagging, NER with AraBERT and CAMeL Assignment: Perform named entity recognition on sample Arabic text Session 12: Sentiment and Emotion Analysis in Arabic # Sentiment analysis models and techniques Applying pre-trained sentiment models to Arabic datasets Assignment: Build a basic Arabic sentiment classifier Session 13: Web Scraping for Data Collection # Tools overview: BeautifulSoup, Selenium Scraping Arabic web data for NLP preprocessing Assignment: Write a scraper for Arabic news articles Session 14: Integrating NLP Pipeline with Scraped Data # Pipeline overview: scraping → cleaning → analysis End-to-end example with Arabic text analysis on scraped data Assignment: Complete a mini pipeline project with hints Session 15: Level 3 Review and Project Assignment # Review Arabic NLP, scraping basics Project: Analyze sentiments and entities in Arabic social media posts Deliverables and guidelines Level 4: Image Processing and Emotion Detection (5 Sessions) # Session 16: Image Processing Basics with OpenCV # Image tasks: resizing, filtering, color spaces Hands-on OpenCV scripts Assignment: Build an image preprocessor Session 17: Object Detection with YOLO # YOLO architecture and inference Running YOLO models on custom images/videos Assignment: Implement object detection and generate reports Session 18: Emotion and Sentiment Detection from Images # Emotion datasets and models (FER, DeepFace) Sentiment analysis from facial expressions Assignment: Build a simple emotion detector on sample images Session 19: Integrating Image Processing and NLP Results # Multimodal analysis combining text and images Example: Analyze social media posts with text and image sentiment Assignment: Build a multimodal mini-project Session 20: Level 4 Review and Project Assignment # Recap image processing, detection, sentiment analysis Project: Social media post analyzer combining image and text emotion detection Assessment criteria Level 5: Neural Networks and Advanced ML (5-7 Sessions) # Session 21: Neural Network Fundamentals # Perceptrons, activation functions, forward pass Simple neural net example in TensorFlow/PyTorch Assignment: Build a perceptron for binary classification Session 22: Loss Functions and Backpropagation # MSE, cross-entropy explained Backpropagation illustrated with examples Assignment: Implement custom loss function Session 23: Supervised Learning Techniques # Linear/logistic regression, decision trees overview Training and evaluation practice Assignment: Build a supervised classifier from scratch Session 24: Unsupervised Learning Techniques # Clustering, dimensionality reduction Visualization with PCA, t-SNE Assignment: Cluster an unlabeled dataset Session 25: Deep Learning Architectures # CNNs, RNNs, LSTMs overview Transfer learning with pretrained models Assignment: Fine-tune a CNN on a sample dataset Session 26: Pros and Cons of Algorithms, Use Cases # Comparative analysis of algorithms When to use supervised vs unsupervised, deep vs shallow Assignment: Choose algorithms for problem statements and justify Session 27: Final Course Review and Graduation Project Introduction # Review core concepts from all levels Present graduation project guidelines: e.g., Arabic text and image sentiment app with deployment Q\u0026amp;A and next steps ","externalUrl":null,"permalink":"/courses/mlops-from-zero-to-full-stack-ai-engineer/","section":"Welcome to Your Data Science \u0026 AI Journey","summary":"27-session AI/ML journey from chatbot APIs and local models to Arabic NLP, multimodal AI, deep learning, and production deployment.","title":"MLops - from Zero to Full Stack AI Engineer","type":"courses"},{"content":" Open Source Contributions # Contributing to major open source projects in computer vision, AI, and robotics.\nContribution Statistics # Repository PRs Merged PRs Open Approved OpenCV (opencv/opencv) 2+ 9 1 opencv-python 0 4 0 Ultralytics 2 3 0 OpenDR (opendr-eu/opendr) 0 1 1 TOTAL 4+ 17 2 OpenCV Contributions # Overview # OpenCV (Open Source Computer Vision Library) is the world\u0026rsquo;s largest computer vision library with 2500+ optimized algorithms.\nMerged PRs # PR # Issue Description Status #28880 #26496 highgui(qt): preserve UTF-8 window names in fallback paths Merged #28936 #28930 core: fix integer overflow in Point_::dot() Merged Open PRs (OpenCV Core) # PR # Issue Title Status #28935 #21960 core: fix Unicode temp path handling on Windows ✅ Approved #28974 #28926 fix(core): prevent signed overflow in cv::cubeRoot() Open #28975 #21037 fix(core): prevent overflow in Rect::br() Open #28976 #28913 fix(core): prevent intra-object overflow in cv::Mat Open #28977 #28949 fix(core): prevent intra-object underflow in cv::MatSize Open #28978 #28940 fix(imgproc): prevent UB in ThickLine left-shift Open #28979 #28948 fix(core): prevent intra-object overflow in Mat::total() Open #28984 #23577 fix(core): prevent use-after-scope in MatExpr Open #28985 #14475 fix(core): prevent overflow in cvInitImageHeader Open Open PRs (opencv-python) # PR # Issue Title #1222 #993 fix: use setuptools\u0026gt;=68.0.0 for Python 3.12+ #1223 #1201 fix: add NumPy version constraints for Python 3.13+/3.14 #1224 #1191 manylinux: remove bundled OpenSSL to fix FIPS failure #1225 #1166 docs: add multiprocessing fork safety warning Types of Fixes # Integer Overflow Prevention: cv::cubeRoot(), Rect::br(), cv::Mat, Mat::total() Undefined Behavior: ThickLine left-shift operations, use-after-scope Unicode Support: UTF-8 window names on Windows Python Compatibility: NumPy version constraints, setuptools requirements Ultralytics Contributions # Overview # Ultralytics is the creator and maintainer of YOLO - the most popular object detection model.\nOpen PRs # PR # Issue Title #24437 #24063 fix: prevent NaN losses in AdamW training for YOLO26 #24438 #24408 fix: enable YOLO26 Pose training with DDP + torch.compile #24439 #24399 fix: ensure consistent mAP on MPS during training validation Direct Merged Commits # Commit Issue Description 122de82 #24272 Correct process_mask crop order in nn/tasks.py c629e2c #24388 Avoid unnecessary padding in set_rectangle Types of Fixes # Training Stability: NaN loss prevention in AdamW optimizer Distributed Training: DDP + torch.compile compatibility Platform Support: Consistent mAP on Apple MPS OpenDR Contributions # Overview # OpenDR (Open Domain Radiance) is a modular, open-source library for robotic vision and learning.\nPR #522 # Field Details Issue #515 Title fix: Update av dependency to \u0026gt;=12.0.0 Status ✅ Approved (waiting for merge) Approver omichel Personal Open Source Project # opencode-presentations-skill # Field Details Author Mulham Fetna Version 2.0.0 Type OpenCode Plugin GitHub mulhamfetna/opencode-presentations-skill Features:\n20 CSS design styles (Modern, Professional, Creative, Tech) 5 slide sizes (16:9, 4:3, 1:1, 9:16, 21:9) Carousel and chart templates AI image generation via Pollinations.ai Live preview with hot reload Web scraping for design inspiration Full test suite (20 tests, 100% pass) GitHub Profile # Username: mulhamfetna\nMerged contributions: 3 pull requests accepted into OpenCV and OpenDR, from 30\u0026#43; submitted across four projects.\nPull request Project What it fixed opencv#28935 OpenCV (core) Unicode temp-path handling on Windows opencv#28880 OpenCV (highgui/Qt) UTF-8 window names preserved in fallback paths opendr#522 OpenDR Python 3.12 compatibility Two of these landed in OpenCV core. Contributing to a library used by millions means most submissions are rejected — I have opened 30\u0026#43; PRs across OpenCV, opencv-python, Ultralytics, and OpenDR, and these three were accepted.\nContribution Philosophy # \u0026ldquo;Contributing to open source is about giving back to the community that enables modern software development. Each fix, no matter how small, improves the reliability and performance of software used by millions worldwide.\u0026rdquo;\nFocus Areas # Bug Fixes: Undefined behavior, integer overflows, memory safety Performance: Optimization for edge devices and embedded systems Compatibility: Cross-platform support, new Python versions Documentation: Improving clarity and examples Skills Demonstrated by Contributions # Category Technologies Languages C++, Python, CMake Computer Vision OpenCV, image processing Deep Learning YOLO, PyTorch, torch.compile Debugging ASAN, UBSAN, undefined behavior Build Systems CMake, setuptools, manylinux Cross-Platform Windows, Linux, macOS ","externalUrl":null,"permalink":"/open-source/","section":"Open Source Contributions","summary":"Open Source Contributions # Contributing to major open source projects in computer vision, AI, and robotics.\nContribution Statistics # Repository PRs Merged PRs Open Approved OpenCV (opencv/opencv) 2+ 9 1 opencv-python 0 4 0 Ultralytics 2 3 0 OpenDR (opendr-eu/opendr) 0 1 1 TOTAL 4+ 17 2 OpenCV Contributions # Overview # OpenCV (Open Source Computer Vision Library) is the world’s largest computer vision library with 2500+ optimized algorithms.\nMerged PRs # PR # Issue Description Status #28880 #26496 highgui(qt): preserve UTF-8 window names in fallback paths Merged #28936 #28930 core: fix integer overflow in Point_::dot() Merged Open PRs (OpenCV Core) # PR # Issue Title Status #28935 #21960 core: fix Unicode temp path handling on Windows ✅ Approved #28974 #28926 fix(core): prevent signed overflow in cv::cubeRoot() Open #28975 #21037 fix(core): prevent overflow in Rect::br() Open #28976 #28913 fix(core): prevent intra-object overflow in cv::Mat Open #28977 #28949 fix(core): prevent intra-object underflow in cv::MatSize Open #28978 #28940 fix(imgproc): prevent UB in ThickLine left-shift Open #28979 #28948 fix(core): prevent intra-object overflow in Mat::total() Open #28984 #23577 fix(core): prevent use-after-scope in MatExpr Open #28985 #14475 fix(core): prevent overflow in cvInitImageHeader Open Open PRs (opencv-python) # PR # Issue Title #1222 #993 fix: use setuptools\u003e=68.0.0 for Python 3.12+ #1223 #1201 fix: add NumPy version constraints for Python 3.13+/3.14 #1224 #1191 manylinux: remove bundled OpenSSL to fix FIPS failure #1225 #1166 docs: add multiprocessing fork safety warning Types of Fixes # Integer Overflow Prevention: cv::cubeRoot(), Rect::br(), cv::Mat, Mat::total() Undefined Behavior: ThickLine left-shift operations, use-after-scope Unicode Support: UTF-8 window names on Windows Python Compatibility: NumPy version constraints, setuptools requirements Ultralytics Contributions # Overview # Ultralytics is the creator and maintainer of YOLO - the most popular object detection model.\n","title":"Open Source Contributions","type":"open-source"},{"content":"Most talented people are invisible. Not because they lack the work — because nobody can find it, verify it, or understand it in ten seconds.\nThis workshop fixes that. It is the same programme I built from scratch and have delivered to 270 participants online and 75 in person at Homs University\u0026rsquo;s IT Department in 2026.\nFormat # 9 hours — 3 sessions × 3 hours.\nSession 1 — interactive dialogue. We start with your actual situation, not slides. What are you trying to be found for, and by whom? Sessions 2 \u0026amp; 3 — directed training. Hands-on building, with your real profile and your real CV open in front of you. Delivered in Arabic and English, online or in person.\nWhat you build # Your positioning. Who you are professionally, in one sentence someone else can repeat. Most people cannot do this, and it is why their CV reads like everyone else\u0026rsquo;s.\nYour CV and résumé. Structure, wording, and the difference between listing duties and evidencing outcomes. What a reviewer actually reads, and in what order.\nYour online presence. Profile optimization — headline, sections, keywords, and the search behaviour that determines whether you appear at all. Professional photography and banner basics.\nYour content and network. How to write posts that build authority rather than noise, how to engage so that people remember you, and how to build relationships that outlast a single application.\nWho it\u0026rsquo;s for # University students and recent graduates entering a crowded market, engineers and professionals who do good work nobody sees, and researchers who need a discoverable academic profile.\nNo prior experience needed. You leave with the assets built, not with a to-do list.\nFor researchers and academics # A dedicated variant covers researcher visibility: ORCID, Google Scholar, ResearchGate, citable software with DOIs, and open-access publishing — building a research profile that a committee or collaborator can actually find and verify. This is the material I teach to researchers in the Scientific Research Camp.\nBook this workshop # Available for universities, student organizations, and companies. Get in touch or email contact@mulhamfetna.com.\nCommon questions # How long is the Personal Branding 101 workshop? Nine hours, delivered as three sessions of three hours each. The first is an interactive dialogue session; the second and third are directed, hands-on training. Is the personal branding workshop available in Arabic? Yes. It is delivered in Arabic and English, online or in person. Who has taken this workshop? Two cohorts in 2026: 270 participants online through Boundless, and 75 in person at the IT Department of Homs University. It is designed for students, graduates, engineers, and researchers. Does it cover academic and research profiles? Yes. A dedicated variant covers researcher visibility — ORCID, Google Scholar, ResearchGate, citable software with DOIs, and open-access publishing. ","externalUrl":null,"permalink":"/workshops-camps/personal-branding-101/","section":"Workshops \u0026 Camps","summary":"Your Zero to Hero Professional Presence Guide","title":"Personal Branding 101","type":"workshops-camps"},{"content":"","externalUrl":null,"permalink":"/projects/","section":"Projects","summary":"","title":"Projects","type":"projects"},{"content":" Python Data Engineering \u0026amp; MLOps Comprehensive Roadmap # This is a high-detail, execution-first roadmap designed to move you from beginner-to-professional in data engineering, machine learning, deep learning, and MLOps.\nThis roadmap is tightly aligned with:\nPython from Zero to Data Engineering Mastery MLops - from Zero to Full Stack AI Engineer Use this roadmap as your sequencing engine and use both course pages as implementation blueprints.\nTo execute this roadmap with live guidance and practical announcements, combine it with:\nWorkshops \u0026amp; Camps Mentorship Services Quick Navigation # Learning philosophy and execution model Phase roadmap (0-48 weeks) Specialization tracks and capstones Tool stack and role-mapped pathways Project difficulty matrix FAQ Appendix (bilingual metadata, job/title landscape, synthetic data resources) Learning Philosophy and Execution Model # Core philosophy # Depth over breadth: master a focused stack deeply. Portfolio over passive learning: every phase ends with an artifact. Production mindset from early stages: testing, versioning, reproducibility. Business relevance: tie every project to a real decision, KPI, or workflow. Minimum execution rule per phase # For each phase, complete:\nOne learning sprint from this roadmap. One implementation using one of the two course pages. One portfolio artifact (repo/notebook/API/dashboard/demo). Recommended weekly routine (high-intensity track) # Coding and project work: 10-20 hours Theory and reading: 4-6 hours Practice tasks (SQL/ML): 3-5 hours Reflection and documentation: 2 hours Community/networking: 1 hour Program Timeline and Course Mapping # Suggested route for beginners # Start with Python from Zero to Data Engineering Mastery for foundation, programming maturity, and data workflows. Then run MLops - from Zero to Full Stack AI Engineer for AI systems, multimodal workflows, and production deployment. Suggested route for experienced Python learners # Run this roadmap phases in order. Use Python course for reinforcement gaps (SQL/data engineering/core systems). Use MLOps course for advanced ML/NLP/CV/deployment progression. Phase 0: Environment, Workflow, and Setup (Day 0 to Week 0) # Objective # Create a reliable local development environment with reproducible workflows.\nChecklist # Linux development environment ready. Git and GitHub configured. Python 3.11+ environment strategy (venv/conda/pyenv). IDE setup (VS Code or Neovim). Basic CLI comfort. Baseline setup # pip install pandas numpy matplotlib seaborn jupyter black ruff pytest pip install sqlalchemy duckdb pip install scikit-learn xgboost lightgbm pip install python-dotenv Course tie-in # Foundation habits are reinforced through Python from Zero to Data Engineering Mastery. Phase 1: Python Foundations for Data (Weeks 1-4) # Objective # Write idiomatic Python for data-heavy tasks and modular pipelines.\nCore topics # Functions, classes, scope, exceptions, logging. Type hints, script structure, reusable modules. NumPy array fundamentals, vectorization, broadcasting. Pandas dataframes, cleaning, joins, groupby, datetime, memory optimization. File and format handling (CSV, JSON, SQL-ready outputs). Practice expectations # Complete 3-5 mini data scripts. Build one CLI data utility. Standardize formatting/linting and README habits. Capstone (Phase 1) # Robotics Sensor Data Pipeline\nIngest raw sensor files, clean anomalies, resample, compute rolling stats, and export quality-controlled outputs.\nCourse tie-in # Sessions 1-16 from Python from Zero to Data Engineering Mastery Phase 1.5: Business Tools Layer (Weeks 5-6) # Objective # Bridge technical data work with real business analytics environments.\nWeek 5 topics # Excel/Sheets: pivots, lookup functions, cleaning patterns. Power BI / Looker Studio dashboard fundamentals. AppSheet awareness for no-code operational contexts. Decision boundary: when dashboards are enough vs when pipelines are needed. Week 6 topics # SQL essentials: SELECT, WHERE, ORDER BY, LIMIT. Joins (INNER/LEFT), GROUP BY aggregations. Subqueries and clean aliasing. DuckDB local SQL workflow. Capstone (Phase 1.5) # Business Dashboard from Scratch\nUse a public dataset. Clean with Python/Pandas. Query with DuckDB SQL. Build BI dashboard. Deliver one-page business recommendations brief. Course tie-in # Use data prep and automation workflow from Python from Zero to Data Engineering Mastery. Phase 2: Data Visualization and EDA (Weeks 5-6, parallel reinforcement) # Objective # Turn raw data into trustworthy visual narratives and exploratory insights.\nTopics # Matplotlib object model and publication-quality plots. Seaborn statistical charts and relationship analysis. Correlation maps, distribution diagnostics, trend decomposition. Optional interactive layer with Plotly/Altair. Capstone (Phase 2) # Interactive EDA report with reproducible narrative and visualization exports.\nCourse tie-in # Visualization and dashboard construction from Python from Zero to Data Engineering Mastery. Phase 3: SQL and Data Engineering Foundations (Weeks 7-9) # Objective # Build robust multi-source data pipelines and schema-aware integrations.\nTopics # SQL mastery: joins, CTEs, windows, optimization basics. Python-SQL integration: SQLAlchemy + DuckDB + pandas read_sql/to_sql. ETL patterns, schema standardization, validation checkpoints. Intro orchestration concepts. Capstone (Phase 3) # Multi-source ETL pipeline integrating CSV + DB + API into one validated warehouse dataset.\nCourse tie-in # API, scraping, persistence, and production structuring from Python from Zero to Data Engineering Mastery. Phase 4: Statistical Foundations (Weeks 10-12) # Objective # Build practical statistical reasoning for experimentation and model evaluation.\nTopics # Descriptive statistics and distribution behavior. Probability and core distributions. Confidence intervals and hypothesis testing. A/B test design and interpretation. Effect size and practical significance. Capstone (Phase 4) # A/B testing framework with simulation, analysis, and reusable reporting.\nCourse tie-in # Evaluation logic directly supports model selection and outcomes in MLops - from Zero to Full Stack AI Engineer. Phase 5: Machine Learning Core (Weeks 13-20) # Objective # Build reliable predictive systems with robust evaluation and tuning.\nTopics # ML workflow lifecycle: framing → splitting → training → validation. Feature scaling and categorical encoding strategies. Supervised learning: Classification: Logistic Regression, Trees/Random Forests, KNN, Naive Bayes, SVM (conceptual depth) Regression: Linear/Ridge/Lasso/Elastic Net, polynomial features Hyperparameter optimization: GridSearch, RandomizedSearch, Bayesian approaches Gradient boosting: XGBoost, LightGBM, CatBoost Imbalanced learning: class weighting, threshold tuning, SMOTE Unsupervised learning: K-Means, DBSCAN, hierarchical clustering, PCA/t-SNE/UMAP Pipeline engineering: Pipeline, ColumnTransformer, custom transformers, joblib Capstone (Phase 5) # Predictive maintenance system with comparative models, tuning, and reproducible evaluation.\nCourse tie-in # Core algorithm and evaluation modules from MLops - from Zero to Full Stack AI Engineer. Phase 6: Deep Learning (Weeks 21-28) # Objective # Move from classical ML to modern neural architectures with practical deployment readiness.\nTopics # Neural network fundamentals: activations, losses, backprop, optimization PyTorch workflow: tensors, autograd, modules, dataloaders, training loops, checkpointing Computer vision: CNN progression, transfer learning, augmentation, YOLO fundamentals NLP foundations: preprocessing, embeddings, sequence models, transformer introduction Hugging Face ecosystem and model usage patterns Capstone options (Phase 6) # Robot vision API (YOLO + FastAPI) Arabic/technical text analysis system (transformers + retrieval/classification) Course tie-in # NLP, CV, multimodal, and advanced model sections in MLops - from Zero to Full Stack AI Engineer. Phase 7: MLOps and Production Systems (Weeks 29-36) # Objective # Ship, monitor, and iterate production ML systems.\nTopics # Experiment tracking and model registry (MLflow/W\u0026amp;B concepts). Model serving with FastAPI/Flask. Serialization formats and deployment tradeoffs. Containerization with Docker + docker-compose. CI/CD pipelines with GitHub Actions. Monitoring and drift awareness. Data versioning and orchestration fundamentals (DVC/Airflow/Prefect concepts). Capstone (Phase 7) # End-to-end ML platform train → validate → register → serve → monitor → retrain loop.\nCourse tie-in # Use deployment and systems modules from both: Python from Zero to Data Engineering Mastery MLops - from Zero to Full Stack AI Engineer Phase 8: Specialization, Integration, and Final Capstone (Weeks 37-48) # Objective # Choose one professional track and deliver one portfolio centerpiece.\nTrack A: Computer Vision and Multimodal AI # Augmentation pipelines. Custom YOLO training. Face/emotion analysis. Image-text multimodal pipelines. Real-time inference deployment. Track B: NLP and Arabic AI # Arabic tokenization/morphology/dialect handling. Arabic preprocessing pipelines. AraBERT/CAMeL/AraGPT model workflows. Arabic retrieval/classification systems. Arabic RAG systems. Track C: MLOps and AI Infrastructure # Full ML CI/CD lifecycle. Feature-store concepts. Drift detection and reliability observability. Kubernetes concepts for ML. Cost/performance optimization decisions. Track D: Generative AI Product Engineering # LoRA/QLoRA fundamentals. RAG architecture from ingestion to generation. Agent/tool orchestration patterns. Prompt/evaluation systems. Product reliability beyond demos. Final capstone requirements # Solve a real non-toy problem. Use at least two data sources. Include proper data pipeline. Deploy model/AI feature. Include monitoring and alerting. Include complete documentation and reproducibility. Deliver deployed app + repository + short walkthrough video + technical write-up. Course tie-in # Production and specialization implementation from MLops - from Zero to Full Stack AI Engineer Engineering discipline and systems foundation from Python from Zero to Data Engineering Mastery Project Difficulty Matrix # Project Phase Difficulty Skills Used Estimated Time CSV cleaner + summary stats CLI 1 ⭐ Beginner Python, Pandas, argparse 1-2 days Business dashboard from open data 1.5 ⭐ Beginner SQL, Power BI/Looker, Pandas 3-4 days Full EDA notebook with narrative 2 ⭐ Beginner Pandas, Matplotlib, Seaborn 2-3 days Multi-source ETL pipeline 3 ⭐⭐ Intermediate SQL, DuckDB, Python, validation 1 week A/B test simulation framework 4 ⭐⭐ Intermediate Statistics, SciPy, Pandas 1 week End-to-end classification or regression 5 ⭐⭐ Intermediate scikit-learn, pipelines, evaluation 1-2 weeks Predictive maintenance with sensor data 5 ⭐⭐⭐ Advanced ML, feature engineering, time series 2 weeks Custom image classifier with transfer learning 6 ⭐⭐⭐ Advanced PyTorch, TorchVision, fine-tuning 2 weeks NLP classification or sentiment pipeline 6 ⭐⭐⭐ Advanced Transformers, Hugging Face, evaluation 2 weeks Deployed ML API with monitoring 7 ⭐⭐⭐⭐ Expert FastAPI, Docker, MLflow, CI/CD 3 weeks Full RAG chatbot with real documents 8 ⭐⭐⭐⭐ Expert LangChain, vector DB, LLM APIs 2-3 weeks Arabic NLP pipeline end-to-end 8 ⭐⭐⭐⭐ Expert Arabic models, preprocessing, deployment 3 weeks Real-time object detection API 8 ⭐⭐⭐⭐ Expert YOLOv8, OpenCV, FastAPI, Docker 3 weeks Rule: always choose one active project slightly above current comfort zone.\nComplete Tool Stack (Consolidated) # Core development # Python 3.11+ Jupyter VS Code + extensions Git + GitHub Black + Ruff pytest Data stack # NumPy Pandas DuckDB SQLAlchemy Optional Polars (later) Visualization stack # Matplotlib Seaborn Plotly (interactive layer) Machine learning stack # scikit-learn XGBoost LightGBM / CatBoost Optuna (optional tuning extension) Deep learning stack # PyTorch TorchVision transformers Hugging Face datasets/tokenizers MLOps stack # MLflow FastAPI Docker DVC Airflow/Prefect concepts Prometheus + Grafana (monitoring concepts) Specialization stack (choose per track) # CV: OpenCV, Albumentations, YOLO NLP: spaCy, Transformers, sentence embeddings GenAI: LangChain/LlamaIndex, vector DBs Infra: CI/CD + observability + cost optimization Role-Target Pathways (Condensed) # Fastest analyst path # Excel/Sheets SQL BI tool (Power BI/Tableau/Looker) Add Python for automation and complex preprocessing Data scientist / ML engineer path (recommended core) # SQL Python data stack scikit-learn PyTorch MLOps fundamentals Data engineer path # SQL + Python pipelines ETL/ELT patterns Orchestration and data quality Deployment and observability FAQ # 1. Should I do Python course before MLOps course? # Yes for most learners. Start with Python from Zero to Data Engineering Mastery then progress to MLops - from Zero to Full Stack AI Engineer.\n2. Can I run both in parallel? # Yes, if Python basics are strong. Use Python course for structure and MLOps course for advanced application.\n3. Which course is best for SQL + data engineering foundations? # Python from Zero to Data Engineering Mastery, then production expansion in MLops - from Zero to Full Stack AI Engineer.\n4. Which phases matter most for AI Engineer roles? # Phases 5-8 plus advanced modules in MLops - from Zero to Full Stack AI Engineer.\n5. Which phases matter most for Data Engineer roles? # Phases 1, 1.5, 3, and 7 with strong execution from Python from Zero to Data Engineering Mastery.\n6. Do I need every tool listed here? # No. Master core stack first, then choose specialization tools based on role target.\n7. What portfolio is minimum viable for job applications? # At least:\nOne clean data pipeline project. One evaluated ML project. One deployed API/dashboard with docs. 8. How do I prepare for interviews from this roadmap? # For every completed phase, prepare:\nArchitecture explanation Tradeoff explanation Reproducible demo 9. Can this roadmap support Arabic NLP specialization? # Yes, via Phase 6 + Phase 8 Track B and MLops - from Zero to Full Stack AI Engineer.\n10. How frequently should I revisit course pages? # Weekly. Treat roadmap as sequence and courses as implementation references.\n11. How much time does the full track take? # Roughly 12-18 months at consistent execution pace.\n12. What if I get stuck in one phase? # Freeze new-tool expansion, complete one scoped project, and only then continue.\n13. Is this roadmap suitable for freelancers? # Yes. Prioritize deployment, reproducibility, and business-facing artifacts.\n14. Is Linux required? # Not strictly required, but Linux-first workflows are strongly recommended for engineering stability.\n15. What is the most common failure pattern? # Consuming tutorials without shipping projects. This roadmap is intentionally project-first to prevent that.\nAppendix A: Bilingual Source Metadata # معلومات المستند | Document Information # العنوان / Title خارطة طريق علم البيانات وتعلم الآلة / Python Data Science \u0026amp; Machine Learning Roadmap النسخة / Version 1.1 (integrated) التاريخ / Date March 2026 المؤلف / Author Eng. Mulham Fetna المسمى الوظيفي / Title CEO \u0026amp; Founder المنظمة / Organization Neurobotics Academy Intellectual property notice (source context) # This integrated roadmap consolidates educational planning materials and preserves attribution context from prior source drafts.\nAppendix B: Job Titles, Tools, and Domain Landscape (Consolidated) # Job family landscape # Core analytics: Data Analyst, BI Analyst, Operations Analyst Data engineering: Data Engineer, Analytics Engineer, Data Platform Engineer Data science: Data Scientist, Applied Scientist, Decision Scientist ML/MLOps: ML Engineer, MLOps Engineer, AI Platform Engineer AI applications: AI Engineer, GenAI Engineer, NLP/CV Engineer Leadership: Analytics Manager, Head of Data/AI, Director roles Core tool families by practical need # Spreadsheet + BI for fast business reporting. SQL for data access and transformation. Python for automation, unstructured data, ML pipelines, and integration. Production stack for deployment, monitoring, and reliability. Appendix C: Synthetic Data and Innovation Monitoring References # Synthetic data references # SDV (Synthetic Data Vault) YData synthetic tooling Gretel.ai Hugging Face synthetic data resources NVIDIA/IBM synthetic data explainers Innovation monitoring references # EU/JRC TIM analytics ecosystem EDPS weak-signal monitoring context OECD AI and policy observatories arXiv and research-trend monitoring Use case note # Use synthetic data and weak-signal monitoring as advanced exploration topics after completing core roadmap execution milestones.\nAppendix D: Practical Career Execution Notes # 30-day starter plan # Week 1: SQL fundamentals and query practice Week 2-3: Python data stack practical projects Week 4: One dashboard + one mini deployment artifact Portfolio quality rule # One excellent documented project is better than multiple shallow tutorial clones.\nFinal guidance # Build in public, document decisions, and keep shipping.\n","externalUrl":null,"permalink":"/roadmaps/python-data-to-mlops-roadmap/","section":"Your Zero-to-Hero Path to Industry-Ready Expertise","summary":"Deep, integrated roadmap from Python foundations to production AI systems with strong alignment to the Python Mastery and MLOps courses.","title":"Python Data Engineering \u0026 MLOps Comprehensive Roadmap","type":"roadmaps"},{"content":" Python Mastery Bootcamp 2026 — Course overview # Neurobotics Academy · Lead instructor: Eng. Mulham Fetna\nMission # Take beginners to production-style Python: fundamentals, data work, databases, dashboards, OOP, Git, scraping, APIs/async, and a graduation capstone (“Data Intelligence Platform” pattern).\nLevel: Beginner → job-ready portfolio (Python + data + shipping).\nSessions: 24 (see SCHEDULE for exact titles and artifacts).\nPacing (per cohort) # Option A: 2 sessions/week → ~12 weeks (~3 months) Option B: 3 sessions/week → ~8 weeks (~2 months) Session curriculum # Phases match SCHEDULE.\nPhase 1 — Foundations (sessions 1–6) # Introduction and Python Setup Python Data Types, Variables, and Basic Arithmetic Conditional Statements \u0026amp; Math Operations Loops + Data Structures + JSON CRUD Functions + Nested Loops + Time Complexity Lists + Modular Functions + Package Intro Phase 2 — Data in Python (sessions 7–11) # Dictionaries + JSON Databases NumPy + Virtual Environments + Statistics Matplotlib + Linear Regression File I/O + OS Navigation + Error Handling Tkinter GUI + Student Attendance System Phase 3 — Databases \u0026amp; analytics (sessions 12–17) # SQL + SQLite + Database Relationships Pandas DataFrames + Functional Programming Advanced Pandas + Export + Streamlit Dashboards Web UI - Markdown + Dash + Streamlit SQLite + Advanced Pandas Data Cleaning Data Pipeline QA, Logging \u0026amp; Reproducible Runs Phase 4 — Software design \u0026amp; shipping (sessions 18–20) # Complete OOP Mastery Functional Programming Mastery Git \u0026amp; GitHub Version Control Mastery Phase 5 — Integration \u0026amp; production (sessions 21–24) # Complete Web Scraping → Production Dashboard Complete REST APIs → Production Data Pipeline Complete Data Science Portfolio Project GRADUATION PROJECT — Complete Python Mastery Showcase Technologies # Python: stdlib, venv, OOP, functional patterns, logging, testing basics Data: NumPy, Pandas, Matplotlib; SQLite; Streamlit / Dash Shipping: Git/GitHub, requirements.txt, structured projects Integration: requests, BeautifulSoup, asyncio / async APIs (as covered in sessions) Graduation deliverable # Capstone integrates scraping/APIs, Pandas processing, SQLite, Streamlit (or similar UI), tests, and GitHub-ready layout. Session 24 defines the full architecture checklist. Fetna.md`\nLicense # Designed for use within Neurobotics Academy. Rights reserved by Neurobotics Academy and the course developer unless otherwise agreed in writing.\nPython Mastery Bootcamp 2026 — session schedule # Single source of truth for order and titles.\nSession Title Primary stack 1 Introduction and Python Setup Python, VS Code 2 Python Data Types, Variables, and Basic Arithmetic Core Python 3 Conditional Statements \u0026amp; Math Operations Control flow 4 Loops + Data Structures + JSON CRUD Loops, JSON 5 Functions + Nested Loops + Time Complexity Functions 6 Lists + Modular Functions + Package Intro Lists, modules 7 Dictionaries + JSON Databases dict, JSON IO 8 NumPy + Virtual Environments + Statistics NumPy, venv 9 Matplotlib + Linear Regression Matplotlib, sklearn intro 10 File I/O + OS Navigation + Error Handling pathlib, exceptions 11 Tkinter GUI + Student Attendance System Tkinter, JSON 12 SQL + SQLite + Database Relationships SQL, SQLite 13 Pandas DataFrames + Functional Programming Pandas, FP 14 Advanced Pandas + Export + Streamlit Dashboards Pandas, Streamlit 15 Web UI - Markdown + Dash + Streamlit Dash, Streamlit 16 SQLite + Advanced Pandas Data Cleaning SQLite, Pandas ETL 17 Data Pipeline QA, Logging \u0026amp; Reproducible Runs logging, validation 18 Complete OOP Mastery OOP, dataclasses 19 Functional Programming Mastery map/filter/reduce 20 Git \u0026amp; GitHub Version Control Mastery Git 21 Complete Web Scraping → Production Dashboard BeautifulSoup, requests 22 Complete REST APIs → Production Data Pipeline requests, asyncio 23 Complete Data Science Portfolio Project Capstone prep 24 GRADUATION PROJECT - Complete Python Mastery Showcase Full-stack capstone Pacing (sper cohort) # 24 sessions total. Option A: 2 sessions/week → ~12 weeks (~3 months) Option B: 3 sessions/week → ~8 weeks (~2 months) Phase grouping # Aligned with the table above (not legacy overview docs).\nFoundations (1–6): setup through lists and modular code Data in Python (7–11): dicts/JSON, NumPy, visualization, files, GUI Databases \u0026amp; analytics (12–17): SQL/SQLite, Pandas, Streamlit/Dash, ETL-style SQLite+Pandas, pipeline QA Software design \u0026amp; shipping (18–20): OOP, functional style, Git/GitHub Integration \u0026amp; production (21–24): scraping, APIs/async, portfolio project, graduation platform ","externalUrl":null,"permalink":"/courses/python-from-zero-to-data-engineering/","section":"Welcome to Your Data Science \u0026 AI Journey","summary":"Beginner-to-professional Python program covering fundamentals, data stack, APIs, async systems, and production portfolio delivery.","title":"Python from Zero to Data Engineering Mastery","type":"courses"},{"content":"","externalUrl":null,"permalink":"/series/","section":"Series","summary":"","title":"Series","type":"series"},{"content":" Technical Skills \u0026amp; Expertise # Comprehensive overview of technical competencies across AI/ML, embedded systems, robotics, and software development.\nProgramming Languages # Language Proficiency Primary Use Python Advanced AI/ML, Data Engineering, Automation, Web C++ Advanced Embedded Systems, OpenCV, Performance-critical C Intermediate Microcontroller Programming C# Intermediate Windows Development Shell/Bash Intermediate Automation, System Administration AI \u0026amp; Machine Learning # Skill Description Tools/Frameworks Deep Learning Neural network development PyTorch, TensorFlow Computer Vision Image processing, object detection OpenCV, YOLO, Ultralytics Edge AI Real-time inference on devices Raspberry Pi, TensorFlow Lite RAG Pipelines Retrieval Augmented Generation LangChain, Ollama, LLaMA MLOps ML operations \u0026amp; deployment Docker, MLflow, CI/CD Arabic NLP Arabic text processing AraBERT, pyarabic, CAMeL Tools Model Training Custom model fine-tuning PyTorch, TensorFlow Local AI On-premise LLM deployment Ollama, LLaMA, FastAPI Robotics \u0026amp; Control Systems # Skill Description Application ROS/ROS2 Robot Operating System Robotics frameworks, Gazebo Model Predictive Control MPC for UAVs Multi-UAV formation Geometric Control SO(3) attitude control Quaternion-based Inverse Kinematics Hybrid ensemble IK Robotic manipulation Path Planning Swarm algorithms Dynamic environments State Estimation EKF/UKF Localization RTOS Real-time operating systems Embedded Control Theory PID, LQR, Optimal Control systems Embedded Systems \u0026amp; Hardware # Skill Description Tools Microcontrollers ATmega, STM32, ESP32 Arduino, PlatformIO PCB Design Circuit design KiCad, EasyEDA, Proteus Firmware Development Embedded programming C, C++ Sensors IR, ultrasonic, IMU I2C, SPI, UART Motors Brushed, brushless, stepper, servo PWM control Communication I2C, SPI, UART, CAN, Bluetooth, WiFi Embedded protocols Data Engineering # Skill Description Tools Data Analysis Data manipulation Pandas, NumPy Visualization Charts and dashboards Plotly, Dash, Streamlit Parallel Computing Large-scale processing Dask Caching Data caching Redis Databases SQL \u0026amp; NoSQL SQLite, MongoDB Time-Series Financial data analysis Python, NumPy DevOps \u0026amp; Infrastructure # Skill Description Tools Containerization App packaging Docker, Docker Compose Automation Workflow automation N8N Secure Access Remote access Twingate, Cloudflare Tunnel Linux System administration Ubuntu, Arch Version Control Code management Git, GitHub, GitLab CI/CD Continuous integration GitHub Actions Web \u0026amp; API Development # Skill Description Tools REST APIs API development FastAPI, Flask Web Scraping Data extraction BeautifulSoup, Selenium MCP Servers Model Context Protocol Custom implementations API Integration Third-party APIs OpenAI, Google, Anthropic Documentation \u0026amp; Tools # Skill Description Tools Academic Writing Research papers LaTeX Technical Docs Documentation Markdown, Mermaid Diagrams Visual documentation Mermaid, diagrams.net Mathematics \u0026amp; Statistics # Skill Description Probability Theory Advanced Statistics Advanced Linear Algebra Matrices, transformations Time-Series Analysis Financial data Quaternion Representations Robotics Optimization Metaheuristics Languages # Language Proficiency Arabic Native English Fluent Skill Levels Summary # pie title Technical Skills Distribution \u0026#34;AI/ML \u0026amp; Data\u0026#34; : 30 \u0026#34;Robotics \u0026amp; Control\u0026#34; : 25 \u0026#34;Embedded Systems\u0026#34; : 20 \u0026#34;Web \u0026amp; API\u0026#34; : 15 \u0026#34;DevOps \u0026amp; Tools\u0026#34; : 10 Certifications # Certification Issuer CS50 Certified Instructor Harvard University McKinsey Forward Program McKinsey \u0026amp; Company Practical Electronics Diploma Analog \u0026amp; Digital Design AWS AI \u0026amp; ML Scholarship AWS Google 2025 AI Intensive Course Google PMP Preparation Professional MEAL Methodology Professional Cornell Take The Lead Cornell University Harvard Aspire Leaders Harvard Business School Tools \u0026amp; Technologies Used # Development Environments # VS Code Jupyter Notebook Google Colab Kaggle Cloud Platforms # AWS Google Cloud Version Control # GitHub GitLab Testing # pytest Custom test frameworks ","externalUrl":null,"permalink":"/skills/","section":"Skills \u0026 Tech Stack","summary":"Technical Skills \u0026 Expertise # Comprehensive overview of technical competencies across AI/ML, embedded systems, robotics, and software development.\nProgramming Languages # Language Proficiency Primary Use Python Advanced AI/ML, Data Engineering, Automation, Web C++ Advanced Embedded Systems, OpenCV, Performance-critical C Intermediate Microcontroller Programming C# Intermediate Windows Development Shell/Bash Intermediate Automation, System Administration AI \u0026 Machine Learning # Skill Description Tools/Frameworks Deep Learning Neural network development PyTorch, TensorFlow Computer Vision Image processing, object detection OpenCV, YOLO, Ultralytics Edge AI Real-time inference on devices Raspberry Pi, TensorFlow Lite RAG Pipelines Retrieval Augmented Generation LangChain, Ollama, LLaMA MLOps ML operations \u0026 deployment Docker, MLflow, CI/CD Arabic NLP Arabic text processing AraBERT, pyarabic, CAMeL Tools Model Training Custom model fine-tuning PyTorch, TensorFlow Local AI On-premise LLM deployment Ollama, LLaMA, FastAPI Robotics \u0026 Control Systems # Skill Description Application ROS/ROS2 Robot Operating System Robotics frameworks, Gazebo Model Predictive Control MPC for UAVs Multi-UAV formation Geometric Control SO(3) attitude control Quaternion-based Inverse Kinematics Hybrid ensemble IK Robotic manipulation Path Planning Swarm algorithms Dynamic environments State Estimation EKF/UKF Localization RTOS Real-time operating systems Embedded Control Theory PID, LQR, Optimal Control systems Embedded Systems \u0026 Hardware # Skill Description Tools Microcontrollers ATmega, STM32, ESP32 Arduino, PlatformIO PCB Design Circuit design KiCad, EasyEDA, Proteus Firmware Development Embedded programming C, C++ Sensors IR, ultrasonic, IMU I2C, SPI, UART Motors Brushed, brushless, stepper, servo PWM control Communication I2C, SPI, UART, CAN, Bluetooth, WiFi Embedded protocols Data Engineering # Skill Description Tools Data Analysis Data manipulation Pandas, NumPy Visualization Charts and dashboards Plotly, Dash, Streamlit Parallel Computing Large-scale processing Dask Caching Data caching Redis Databases SQL \u0026 NoSQL SQLite, MongoDB Time-Series Financial data analysis Python, NumPy DevOps \u0026 Infrastructure # Skill Description Tools Containerization App packaging Docker, Docker Compose Automation Workflow automation N8N Secure Access Remote access Twingate, Cloudflare Tunnel Linux System administration Ubuntu, Arch Version Control Code management Git, GitHub, GitLab CI/CD Continuous integration GitHub Actions Web \u0026 API Development # Skill Description Tools REST APIs API development FastAPI, Flask Web Scraping Data extraction BeautifulSoup, Selenium MCP Servers Model Context Protocol Custom implementations API Integration Third-party APIs OpenAI, Google, Anthropic Documentation \u0026 Tools # Skill Description Tools Academic Writing Research papers LaTeX Technical Docs Documentation Markdown, Mermaid Diagrams Visual documentation Mermaid, diagrams.net Mathematics \u0026 Statistics # Skill Description Probability Theory Advanced Statistics Advanced Linear Algebra Matrices, transformations Time-Series Analysis Financial data Quaternion Representations Robotics Optimization Metaheuristics Languages # Language Proficiency Arabic Native English Fluent Skill Levels Summary # pie title Technical Skills Distribution \"AI/ML \u0026 Data\" : 30 \"Robotics \u0026 Control\" : 25 \"Embedded Systems\" : 20 \"Web \u0026 API\" : 15 \"DevOps \u0026 Tools\" : 10 Certifications # Certification Issuer CS50 Certified Instructor Harvard University McKinsey Forward Program McKinsey \u0026 Company Practical Electronics Diploma Analog \u0026 Digital Design AWS AI \u0026 ML Scholarship AWS Google 2025 AI Intensive Course Google PMP Preparation Professional MEAL Methodology Professional Cornell Take The Lead Cornell University Harvard Aspire Leaders Harvard Business School Tools \u0026 Technologies Used # Development Environments # VS Code Jupyter Notebook Google Colab Kaggle Cloud Platforms # AWS Google Cloud Version Control # GitHub GitLab Testing # pytest Custom test frameworks ","title":"Skills \u0026 Tech Stack","type":"skills"},{"content":"Every project here is public, and every claim links to the code. Where a result is weak or a project is early-stage, it says so.\nFault Diagnosis in PMSM Motors — CNN on Wavelet Scalograms # github.com/mulhamfetna/pmsm-fault-diagnosis-cnn-scalogram · Python · MIT\nDetecting inter-turn stator faults in permanent-magnet synchronous motors by converting raw sensor signals into continuous-wavelet scalograms and classifying them with a CNN.\nTrained on the real KAIST motor dataset (DOI 10.17632/rgn5brrgrn.5) — not synthetic data.\nResults, reported honestly:\nSensor channel Balanced accuracy Vibration @ 25.6 kHz 1.00 Current @ 100 kHz 0.69 The vibration result is excellent. The current-channel result is not, and the repository says so. The dataset contains only four healthy recordings, so the perfect vibration score may not generalize — that caveat is in the README, not buried in a footnote.\n38 unit tests. This is the project I would want to be judged on: real data, a strong result, a weak result, and no hiding of either.\nMerged Open-Source Contributions — OpenCV \u0026amp; OpenDR # Three pull requests accepted into major open-source projects. Two landed in OpenCV core, one of the most widely used computer-vision libraries in the world.\nPull request Project What it fixed opencv#28935 OpenCV (core) Unicode temp-path handling on Windows opencv#28880 OpenCV (highgui/Qt) UTF-8 window names preserved in fallback paths opendr#522 OpenDR Python 3.12 compatibility Both OpenCV fixes concern non-ASCII path and text handling — the class of bug that quietly breaks the library for everyone outside the English-speaking world, and that is easy to miss if you never work in Arabic.\nContributing to a library used by millions means most submissions are rejected. I have opened 30\u0026#43; pull requests across OpenCV, opencv-python, Ultralytics, and OpenDR; these three were accepted.\nWiFi as a Radar Replacement for Automotive SLAM # github.com/mulhamfetna/wifi-radar-slam · Python · AGPL-3.0 Archived on Zenodo — DOI 10.5281/zenodo.21247288\nCan ambient WiFi substitute for radar in automotive localization and mapping? A simulation-first feasibility study using Sionna ray tracing, testing whether the channel-state information available from commodity WiFi carries enough structure for SLAM.\nStatus: active research, feasibility stage. The literature review and simulation infrastructure are complete; there are no results yet. I publish it as it develops rather than after the fact.\nTrading Strategy Finder # github.com/mulhamfetna/trading-strategy-finder · Python · MIT\nA backtesting engine comparing scalping, day, and intraday strategies on NQ futures, with a machine-learning filter layer over signal generation.\nFrozen at tag v1.0.0 with reproducible results: $633.65 net profit, 54.5% win rate, 2.62 profit factor.\nModest, believable numbers on a small account — which is the point. The engineering (versioning, tests, reproducibility) is the deliverable here, not a claim of a trading edge.\nOpenCode Presentations # github.com/mulhamfetna/opencode-presentations-skill · JavaScript · MIT · published on npm\nA Marp-based presentation generator with 20 design styles and live preview. Small, finished, and actually shipped — a working tool anyone can install and use today.\nGeopolymer Design System # github.com/mulhamfetna/geopolymer-design-system · Jupyter · Apache-2.0 · archived on Zenodo with a DOI\nA Gradio application pairing a materials database with a RandomForest predictor and a Nelder–Mead optimizer, to design geopolymer mixes for nuclear-waste adsorption.\nCaveat stated up front: the model is trained on synthetic data, not laboratory measurements. It is a design-exploration tool, and explicitly not a validated substitute for experimental work.\nEmbedded Systems — Ala\u0026rsquo;a Screens # Professional work, 2022–2023. Embedded systems design for P10 DMD LED display applications, including a basketball stadium scoreboard clock and a school-bus information display. Firmware and hardware, shipped and deployed.\nCode: github.com/mulhamfetna · Contact: contact@mulhamfetna.com\n","externalUrl":null,"permalink":"/projects/portfolio/","section":"Projects","summary":"Every project here is public, and every claim links to the code. Where a result is weak or a project is early-stage, it says so.\nFault Diagnosis in PMSM Motors — CNN on Wavelet Scalograms # github.com/mulhamfetna/pmsm-fault-diagnosis-cnn-scalogram · Python · MIT\nDetecting inter-turn stator faults in permanent-magnet synchronous motors by converting raw sensor signals into continuous-wavelet scalograms and classifying them with a CNN.\nTrained on the real KAIST motor dataset (DOI 10.17632/rgn5brrgrn.5) — not synthetic data.\nResults, reported honestly:\nSensor channel Balanced accuracy Vibration @ 25.6 kHz 1.00 Current @ 100 kHz 0.69 The vibration result is excellent. The current-channel result is not, and the repository says so. The dataset contains only four healthy recordings, so the perfect vibration score may not generalize — that caveat is in the README, not buried in a footnote.\n38 unit tests. This is the project I would want to be judged on: real data, a strong result, a weak result, and no hiding of either.\nMerged Open-Source Contributions — OpenCV \u0026 OpenDR # Three pull requests accepted into major open-source projects. Two landed in OpenCV core, one of the most widely used computer-vision libraries in the world.\n","title":"Technical Portfolio","type":"projects"},{"content":" Testimonials \u0026amp; Success Stories # Feedback from students, colleagues, and partner organizations.\nStudent Success # Olympic Medal Results (Syrian Scientific Olympiad 2024) # As a coach for the Kids and Adolescents Programming Marathon (KPM):\nGroup Gold Silver Bronze Total Younger 4 10 15 29 Adolescent 8 11 17 36 TOTAL 12 21 32 65 Paper Airplanes - Women in Tech Program # \u0026ldquo;Mulham is an exceptional instructor who transforms complex topics into digestible lessons. His patience and supportive approach helped 26 women from 6 countries transition into tech careers.\u0026rdquo;\nRole: Python Coach \u0026amp; AI Program Mentor (2023-Present)\nProfessional Recognition # Volunteer of the Month — Fall 2025 # Organization: Paper Airplanes\nRecognition for exceptional contribution to Women in Tech Program\nTop 3 — Ramadan Initiatives Award 2026 # Organization: Directorate of Development\nAwarded for excellence in educational services through Neurobotics Academy and Boundless Elite Club\nPartner Organizations # HerWill (US-based women in STEM) # \u0026ldquo;Mulham designed and delivered a comprehensive Python data engineering curriculum for our Women in Tech program. His curriculum is now the standard for our international cohorts.\u0026rdquo;\nBoundless Academic Services # Partner: SANAD and UNFPA Program: Scientific Research Camp (SRC1) Result: 75 enrolled, 35\u0026#43; graduated, 1 publication under review\nAcademic Achievements # Students Mentored: 200+ across 6 countries Courses Delivered: 7+ cohorts (Python, Data Engineering, MLOps) Neurobotics Academy: 50+ graduates Research Camp: 75 first-cohort graduates Testimonials Collection # \u0026ldquo;Eng. Mulham\u0026rsquo;s teaching methodology is unique - he makes complex concepts accessible while maintaining technical rigor. His students consistently achieve top results.\u0026rdquo; — Neurobotics Academy Student\n\u0026ldquo;The Personal Branding 101 workshop changed how I present myself professionally. Highly recommended for anyone serious about their career.\u0026rdquo; — Boundless Workshop Participant\n\u0026ldquo;The Python course exceeded my expectations. From zero knowledge to building my first data pipeline in 3 months.\u0026rdquo; — HerWill Program Graduate\nStatistics at a Glance # Metric Value Students Mentored 200+ Countries Reached 6 Courses Delivered 7+ Olympic Medals (as Coach) 65 Graduation Projects 20+ ","externalUrl":null,"permalink":"/testimonials/","section":"Testimonials","summary":"Testimonials \u0026 Success Stories # Feedback from students, colleagues, and partner organizations.\nStudent Success # Olympic Medal Results (Syrian Scientific Olympiad 2024) # As a coach for the Kids and Adolescents Programming Marathon (KPM):\nGroup Gold Silver Bronze Total Younger 4 10 15 29 Adolescent 8 11 17 36 TOTAL 12 21 32 65 Paper Airplanes - Women in Tech Program # “Mulham is an exceptional instructor who transforms complex topics into digestible lessons. His patience and supportive approach helped 26 women from 6 countries transition into tech careers.”\nRole: Python Coach \u0026 AI Program Mentor (2023-Present)\nProfessional Recognition # Volunteer of the Month — Fall 2025 # Organization: Paper Airplanes\nRecognition for exceptional contribution to Women in Tech Program\nTop 3 — Ramadan Initiatives Award 2026 # Organization: Directorate of Development\n","title":"Testimonials","type":"testimonials"},{"content":"I founded and lead two Syrian ventures. Both exist to open doors that are otherwise closed to people here.\nNeurobotics — Integrated Technology Solutions # Engineering minds and machines.\nNeurobotics Academy — paid diplomas in Python, data engineering, and MLOps for Syrian learners, assessed on attendance and graduation projects. Engineering Team — custom mechatronics and robotics solutions. Neurobotics Junior — STEM labs for school-age students. ROS Arabic Community — the Arabic-language community for Robot Operating System developers, which I founded and manage. Recognized with the Top 3 Ramadan Initiative Award 2026 from the Directorate of Development for excellence in educational services.\nBoundless — Academic Services # Removing barriers to global academic excellence for Syrian and Arab students.\nThe Scientific Research Camp is our flagship: 40 training hours across 11 workshops, delivered with SANAD as strategic partner and UNFPA Syria as donor.\nCamp I (2025) — 75 researchers enrolled, 35\u0026#43; graduated, across engineering, medical, and agricultural research tracks. One publication under review. Camp II (2026) — running now with 50 youth researchers and academics. Boundless also runs scholarship mentorship (three confirmed UWC acceptances), international olympiad training, and academic personal-branding programmes.\nPartners # Partner Organizations الجامعة الافتراضية السورية - Syrian Virtual University Trained \u0026#43;350 mentees for Kids and Adolescents Programming Marathon - One of Only 20 Trainers Nationally\nSyria, Online هيئة التميز والإبداع - Distinction and Creativity Agency Trained \u0026#43;350 mentees for Kids and Adolescents Programming Marathon - One of Only 20 Trainers Nationally\nSyria, Online Kids and Adolescents Programming Marathon - KPM - الماراثون البرمجي للصغار واليافعين Trained \u0026#43;350 mentees for Kids and Adolescents Programming Marathon - One of Only 20 Trainers Nationally\nSyria, Online الأولمبياد العلمي السوري - Syrian Science Olympiad Trained \u0026#43;350 mentees for Kids and Adolescents Programming Marathon - One of Only 20 Trainers Nationally\nSyria, Online Paper Airplanes - Women in Tech Program Trained \u0026#43;60 women through python level 1 (core programming concepts), python level 2 (data engineering) and ai (full stack ai development) courses - Recognized as Mentor of the Month for Fall 2025\nRemote Boundless - Scientific Research Camp Founded and executed a 40-hour scientific research camp with 75 enrolled and 35\u0026#43; graduated — Aleppo University graduates, master\u0026#39;s and doctoral students — with 5 lecturers, focused on applied research. One publication under review.\nSyria, Aleppo HerWILL - Data Engineering Curriculum Designer Designed and delivered a full comprehensive python data engineering curriculum according to our volunteering contract\nRemote Work with us # Sponsorship, partnership, or programme delivery — get in touch, or email contact@mulhamfetna.com.\nCommon questions # What is Neurobotics? Neurobotics is a Syrian engineering and technical-education venture founded by Mulham Fetna. It runs Neurobotics Academy (paid diplomas in Python, data engineering, and MLOps), an engineering team building custom mechatronics and robotics solutions, STEM labs for school students, and the Arabic-language ROS developer community. What is Boundless? Boundless is a Syrian academic-services venture founded by Mulham Fetna, aimed at removing barriers to global academic opportunity for Syrian and Arab students. Its flagship is the Scientific Research Camp; it also runs scholarship mentorship, international olympiad training, and academic personal-branding programmes. What is the Scientific Research Camp? A research-methodology training programme run by Boundless with SANAD as strategic partner and UNFPA Syria as donor. It comprises 40 training hours across 11 workshops. The first cohort (2025) enrolled 75 researchers, of whom 35+ graduated; the second (2026) is running with 50 youth researchers and academics. Who has partnered with these ventures? SANAD and UNFPA Syria back the Scientific Research Camp. Training and educational partnerships include the Syrian Virtual University, the Distinction and Creativity Agency, Paper Airplanes, HerWill, and Homs University. How can I sponsor or partner with these ventures? Email contact@mulhamfetna.com. Both ventures work with donors, universities, companies, and NGOs on programme funding, delivery, and venue or in-kind support. ","externalUrl":null,"permalink":"/ventures/","section":"Ventures","summary":"I founded and lead two Syrian ventures. Both exist to open doors that are otherwise closed to people here.\nNeurobotics — Integrated Technology Solutions # Engineering minds and machines.\nNeurobotics Academy — paid diplomas in Python, data engineering, and MLOps for Syrian learners, assessed on attendance and graduation projects. Engineering Team — custom mechatronics and robotics solutions. Neurobotics Junior — STEM labs for school-age students. ROS Arabic Community — the Arabic-language community for Robot Operating System developers, which I founded and manage. Recognized with the Top 3 Ramadan Initiative Award 2026 from the Directorate of Development for excellence in educational services.\nBoundless — Academic Services # Removing barriers to global academic excellence for Syrian and Arab students.\nThe Scientific Research Camp is our flagship: 40 training hours across 11 workshops, delivered with SANAD as strategic partner and UNFPA Syria as donor.\nCamp I (2025) — 75 researchers enrolled, 35+ graduated, across engineering, medical, and agricultural research tracks. One publication under review. Camp II (2026) — running now with 50 youth researchers and academics. Boundless also runs scholarship mentorship (three confirmed UWC acceptances), international olympiad training, and academic personal-branding programmes.\n","title":"Ventures","type":"ventures"},{"content":" Welcome to Your Data Science \u0026amp; AI Journey # Data science and AI have been the most powerful tools of all time—for the past decade and counting.\nI\u0026rsquo;ve packed two comprehensive courses to supercharge your career: Python from Zero to Data Engineering Course and MLOps from Zero to Full-Stack AI Engineer Course. Let me take you on this transformative journey!\nCheck the organizations and universities I\u0026rsquo;ve managed educational programs with.\nDive into my current courses, key learning subjects, and detailed roadmaps.\nDon\u0026rsquo;t forget to follow my workshops \u0026amp; camps for hands-on mastery.\nStudent Outcomes Snapshot # Suhaila Hussaini: \u0026ldquo;Building a simple app with zero prior knowledge boosted my confidence and made me believe I belong in tech.\u0026rdquo; Salwa Abduljaleel Nagi Ahmed: \u0026ldquo;The course changed my way of thinking and helped me solve work problems more logically and confidently.\u0026rdquo; Ghinwa Allaoui: \u0026ldquo;The coaching style opened doors to practical programming thinking and shortened a long path to real skills.\u0026rdquo; See the full testimonials and learning tracks on the Courses page.\nBuild your full learning path # Use these companion pages while studying:\nRoadmaps hub for full progression plans. Python Data Engineering \u0026amp; MLOps roadmap for data/AI sequencing. Mechatronics Engineering roadmap for robotics/embedded progression. Workshops \u0026amp; Camps for upcoming live sessions and announcements. Mentorship Services for personalized academic/professional guidance. I have managed educational programs with various prestigious institutions: # Partner Organizations الجامعة الافتراضية السورية - Syrian Virtual University Trained \u0026#43;350 mentees for Kids and Adolescents Programming Marathon - One of Only 20 Trainers Nationally\nSyria, Online هيئة التميز والإبداع - Distinction and Creativity Agency Trained \u0026#43;350 mentees for Kids and Adolescents Programming Marathon - One of Only 20 Trainers Nationally\nSyria, Online Kids and Adolescents Programming Marathon - KPM - الماراثون البرمجي للصغار واليافعين Trained \u0026#43;350 mentees for Kids and Adolescents Programming Marathon - One of Only 20 Trainers Nationally\nSyria, Online الأولمبياد العلمي السوري - Syrian Science Olympiad Trained \u0026#43;350 mentees for Kids and Adolescents Programming Marathon - One of Only 20 Trainers Nationally\nSyria, Online Paper Airplanes - Women in Tech Program Trained \u0026#43;60 women through python level 1 (core programming concepts), python level 2 (data engineering) and ai (full stack ai development) courses - Recognized as Mentor of the Month for Fall 2025\nRemote Boundless - Scientific Research Camp Founded and executed a 40-hour scientific research camp with 75 enrolled and 35\u0026#43; graduated — Aleppo University graduates, master\u0026#39;s and doctoral students — with 5 lecturers, focused on applied research. One publication under review.\nSyria, Aleppo HerWILL - Data Engineering Curriculum Designer Designed and delivered a full comprehensive python data engineering curriculum according to our volunteering contract\nRemote My Students Testimonials Speaks for me # Python from Zero to Data Engineering Mastery | Neurobotics Academy | Aleppo Onsite | Q1 2026 # MLops from Zero to Full Stack AI Engineer | Paper Airplanes | Women in Tech | Spring 2026 # Python from Zero to Data Engineering Mastery | Neurobotics Academy | Online | 2025 # Python Level 2 | Paper Airplanes | Women in Tech | Fall 2025 # Suhaila Hussaini\nHi, first of all, I really appreciate the course and your efforts during that. For me, the effect was perhaps indirect. Like when I made a very simple app with zero prior knowledge it boosted my confidence. It increased my experience and knowledge about tech and added one more certificate to my background and got a letter of recommendation which I count as an educational background. Thanks 🙏🏻\nSalwa Abdujaleel Nagi Ahmed\nبصراحة، الكورس غيّر طريقة تفكيري بشكل كبير، وصرت أتعامل مع المشاكل بطريقة أكثر منطقية وثقة. صحيح أني لم أحصل على وظيفة جديدة لحد الآن، لكن الكورس ساعدني كثيراً في تطوير أدائي في عملي الحالي وصرت أفهم النظام بشكل أفضل وحتى قدرت أكتشف بعض الأخطاء فيه وأتعامل معها وحتى بدأت أفكر أشتغل على نظام بديل أعمل عليه يكون أسهل وأوضح، وإن شاء الله أقدر أطبّق الفكرة قريبا بإذن الله. بشكل عام، كان لهذا الكورس تأثير حقيقي في طريقة تفكيري وشغلي، وشكرًا لك على هذا الجهد مدرب ملهم. عن تجربة، أنا درست الكورس مع المدرب ملهم من قبل، والفرق فعلاً واضح. ليس مجرد كورس نظري، بل تطبيق حقيقي بيقرّبك من سوق العمل كل التوفيق إلك مدرب ملهم\nNoor koulayb\nالسلام عليكم ورحمة الله وبركاته عساك بخير إستاذ ملهم. سبحان الله كان عندي حلم اربط بدي لبرمجة والكيمياء صح كان عندي خلفية برمجية ولكن ما كانت تتوافق ما الكيمياء ، مجرد دخولي للكورس بلشت اعرف كيف اربط بين مجالين، كانت خطوة مهمة بالنسبة الي الله يعطيك العافية يا رب. نتمنى نكمل بالمجال ونتوسع اكتر ويصير في متابعة فعلية للمشاريع لتصير علي ارض واقعة ونستفيد منها\nGhinwa Allaoui\nبصراحة،بفضل الكورسات يلي أخدتها مع المدرب ملهم وصلت لمرحلة متطورة بالبرمجة ما كنت متخيلة أوصلها بهالسرعة. والأهم من هيك إنه كمدرّب عم يعطي من قلبه ❤️ وما بيبخل على أي طالب بأي معلومة بالعكس بيفتح كل الأبواب قدامنا من حلول لأفكار لدهاليز ومفاتيح البرمجة يلي عنجد بتختصر طريق طويل. أي حدا عم يفكر يتعلم بايثون أو يفوت بمجال الداتا هيدا المكان الصح 👌\nPython Level 1 | Paper Airplanes | Women in Tech Program | Spring 2025 # Yasmine Ben abdallah\nIt was really nice meeting mr mulham even with the bad connection sometimes but he did he\u0026rsquo;s best to teach us and gave a the resources needed and he was such a great tutor my biggest takeaway from this course is The creativity\nSuhaila Hussaini\nMy biggest takeaway from this course was realizing that I’m capable of building real, useful tools with coding — even with a non-technical background. It gave me the confidence to believe that I belong in tech, and it motivated me to keep learning more and more things in tech. Thanks for your time and energy. You are such a hard-working and kind person. Wish you all the best, and may you achieve all your goals Thanks for this opportunity. For some of us, like me, your projects are life-changing. Due to some circumstances, I could not learn all this in my country, but you made it possible. Thank you once more.\nMira AlHalabi\nmy biggest takeaway I think team work at the last session so it’s developed my soft skills and also my English because it’s my first time to be in a team . Thank you for being near what I ask something also thank you for learning how to think to build a project and how to use AI tools correctly .\nGhinwa Allaoui\nThis course taught me that every problem in life can have multiple solutions, just like in Python, where we break down challenges and find creative ways to solve them to build something useful and meaningful. It changed the way I think and helped me see how coding can simplify and improve real-life tasks. Thank you, Mulham, for your patience and continuous support. Through your guidance and clear teaching, we have opened new doors to knowledge and creativity. You were not just a coach, but a beacon that illuminated our path in programming. Wishing you all the best and much success in your journey. thank you for the incredible opportunity to grow our skills and challenge ourselves. This program was more than just technical learning, it was a journey of empowerment and transformation. We believe this is just the beginning of a bright future for every ambitious woman in technology I applied the course concepts in two small projects: one was a data analysis task at home using Python, where I worked with CSV files, functions, and data visualization. The other was a creative project for kids, I designed a joyful multiplication table with fun animal shapes and favorite characters to help children memorize it in an engaging way. These experiences showed me how coding can be both practical and meaningful.\nSena El emin\nWhen ı attend to these sessions ı wasn\u0026rsquo;t know anything about python. So everything that ı learned from this course was my biggest takeaway. Thank you for everything. I learned a lot from this course. And it will help my carier in the future. Maybe while the course ı asked a lot of questions especially in the final project. So again thank you.\nNarges Saeed\nThe biggest takeaway is how to solve the equation in programming how to convert the equation as code Thanks for your supporting and kindness\nSalwa Abduljaleel Nagi Ahmed\nBefore I started this course, I had no knowledge of the Python language. But throughout the course, I learned the basics step by step. now I feel more confident and I can write simple programs by myself This course helped me believe that I can learn new things, even if they seem difficult at first. I really enjoyed this course and learned a lot from it. The lessons were clear and helpful. For future courses, I suggest adding more practical examples and small projects to help us practice more. Also, maybe you can share short recorded videos for revision after the live sessions. Thank you for your great efforts I want to say a big thank you to Coach Mulham for his great support and help during the course. He was always patient, kind, and encouraging. This really helped me do better in my work and understand the lessons more clearly. I’m very thankful for his time and support. I learned a lot, and this was a very useful and special experience for me. Thank you so much! I would like to express my sincere gratitude to the entire Paper Airplanes – Women in Tech team for this amazing opportunity. This program has been a truly inspiring and empowering experience. I gained not only technical knowledge, but also confidence, motivation, and a sense of community. Special thanks to the mentors, instructors, and organizers for their time, effort, and dedication. Your support made a real difference, and I’m grateful for every moment spent in this program. Thank you for believing in women and creating a space where we can grow, learn, and lead. With appreciation\nKinda Alraiss\nI work in a lab where we need to do weekly samples binning. The samples are binned based on several factors, the type of sample, the date the results released, is it for import or export, there are so many factors. Every time we need to do the binning, we have to calculate the time between the date of results released and the date of binning then see the type of sample, then do the binning. That gives room for mistakes where samples that need to be kept were binned and vice vesa. The application I made allow us to automatically know whether the sample need to be binned or not. Just choose which lab you are doing the binning and what kind of samples. Enter the date of the results released and the application will tell you what to do with the sample.\nSee what past students say: student outcomes and testimonials.\nFeatured Courses \u0026amp; Learning Tracks # ","externalUrl":null,"permalink":"/courses/","section":"Welcome to Your Data Science \u0026 AI Journey","summary":"Welcome to Your Data Science \u0026 AI Journey # Data science and AI have been the most powerful tools of all time—for the past decade and counting.\nI’ve packed two comprehensive courses to supercharge your career: Python from Zero to Data Engineering Course and MLOps from Zero to Full-Stack AI Engineer Course. Let me take you on this transformative journey!\nCheck the organizations and universities I’ve managed educational programs with.\nDive into my current courses, key learning subjects, and detailed roadmaps.\nDon’t forget to follow my workshops \u0026 camps for hands-on mastery.\nStudent Outcomes Snapshot # Suhaila Hussaini: “Building a simple app with zero prior knowledge boosted my confidence and made me believe I belong in tech.” Salwa Abduljaleel Nagi Ahmed: “The course changed my way of thinking and helped me solve work problems more logically and confidently.” Ghinwa Allaoui: “The coaching style opened doors to practical programming thinking and shortened a long path to real skills.” See the full testimonials and learning tracks on the Courses page.\nBuild your full learning path # Use these companion pages while studying:\n","title":"Welcome to Your Data Science \u0026 AI Journey","type":"courses"},{"content":"Mechatronics and AI/ML engineer. I build robotics, computer-vision, and data systems — and I teach them.\nBased in Aleppo, Syria. Available remotely worldwide, in Arabic or English.\nWhat I do # Engineering — robotics and ROS/ROS2, embedded systems (STM32, ESP32, ATmega), control systems, PCB design, edge AI.\nAI, ML \u0026amp; data — machine learning (PyTorch, TensorFlow), computer vision, MLOps, data engineering, and Arabic NLP — a genuinely rare specialism in this region.\nConsulting \u0026amp; mentorship — technical consultation sessions, and one-to-one mentorship for engineers and researchers.\nThe evidence # Don\u0026rsquo;t take the summary — check the work.\nProjects — nine engineering and AI projects, with outcomes. Skills — the full technical competency matrix. Open source — 3 merged pull requests into OpenCV and OpenDR (from 30\u0026#43; submitted), including two fixes accepted into OpenCV core. Full CV — complete experience, affiliations, and certifications. Testimonials — what students and partners say. Get in touch # Book a technical consultation or mentorship session, or email contact@mulhamfetna.com directly.\nCommon questions # Is Mulham Fetna available for hire? Yes. He takes engineering work, technical consulting, and one-to-one mentorship — remotely worldwide, in Arabic or English. He is based in Aleppo, Syria. Contact him at contact@mulhamfetna.com. What are Mulham Fetna\u0026#39;s technical skills? Robotics and ROS/ROS2, embedded systems (STM32, ESP32, ATmega), control systems and PCB design; machine learning with PyTorch and TensorFlow, computer vision, MLOps and data engineering; and Arabic natural language processing. His main languages are Python and C++. What has Mulham Fetna built? Public projects spanning robotics, computer vision, and data systems — including a CNN fault-diagnosis system trained on real motor data — plus 3 pull requests merged into OpenCV and OpenDR, two of them into OpenCV core. Does Mulham Fetna offer consulting or mentorship sessions? Yes. He offers technical consultation sessions and one-to-one mentorship for engineers, students, and researchers, bookable through the Services page. How do I book a session with Mulham Fetna? Book directly through the Services page on this site, or email contact@mulhamfetna.com to discuss a project or engagement. ","externalUrl":null,"permalink":"/work-with-me/","section":"Work With Me","summary":"Mechatronics and AI/ML engineer. I build robotics, computer-vision, and data systems — and I teach them.\nBased in Aleppo, Syria. Available remotely worldwide, in Arabic or English.\nWhat I do # Engineering — robotics and ROS/ROS2, embedded systems (STM32, ESP32, ATmega), control systems, PCB design, edge AI.\nAI, ML \u0026 data — machine learning (PyTorch, TensorFlow), computer vision, MLOps, data engineering, and Arabic NLP — a genuinely rare specialism in this region.\nConsulting \u0026 mentorship — technical consultation sessions, and one-to-one mentorship for engineers and researchers.\nThe evidence # Don’t take the summary — check the work.\nProjects — nine engineering and AI projects, with outcomes. Skills — the full technical competency matrix. Open source — 3 merged pull requests into OpenCV and OpenDR (from 30+ submitted), including two fixes accepted into OpenCV core. Full CV — complete experience, affiliations, and certifications. Testimonials — what students and partners say. Get in touch # Book a technical consultation or mentorship session, or email contact@mulhamfetna.com directly.\n","title":"Work With Me","type":"work-with-me"},{"content":" Learn, Build, and Transform # Welcome to the heart of real-world learning — my workshops, where theory meets action and students evolve into professionals.\nEach workshop is designed to empower you with practical, career-ready skills built on years of leading STEM programs and educational initiatives across organizations and universities.\nIf you’ve ever wondered how data science, AI, robotics, and engineering actually come alive beyond textbooks — this is the place to find out.\nThrough these sessions, we go hands-on with the technologies shaping the present and future of industry, from data pipelines and AI automation to robotics innovation labs and professional growth frameworks. You’ll engage with curated content, collaborative exercises, and engineering-grade challenges that prepare you for serious, real-world projects — not superficial experience.\nFeatured Workshop Tracks # Professional Personal Branding 101\nYour zero-to-hero guide for Gen Z professionals entering the modern workforce. Understand how to brand yourself online and offline with authenticity and strategic purpose.\nData and AI Careers Under the Telescope\nExplore the breadth of data-driven roles — from data analysis and engineering to data science, machine learning, and artificial intelligence. Get insider insights into workflows, tools, and hiring expectations.\nMechatronics \u0026amp; Robotics 101\nStep from fantasy to the factory floor. Discover how mechatronics and robotics power automation, innovation, and research in industries worldwide — blending mechanical theory with digital intelligence.\nEvery workshop connects directly to the Roadmaps and Courses on this site — together forming a full learning ecosystem from zero to industry-ready expertise.\nJoin the next workshop and let’s turn curiosity into capability.\nContinue beyond workshops # Follow the full learning tracks in Courses. Plan long-term progression in Roadmaps. Get one-to-one help through Mentorship Services. ","externalUrl":null,"permalink":"/workshops-camps/","section":"Workshops \u0026 Camps","summary":"Learn, Build, and Transform # Welcome to the heart of real-world learning — my workshops, where theory meets action and students evolve into professionals.\nEach workshop is designed to empower you with practical, career-ready skills built on years of leading STEM programs and educational initiatives across organizations and universities.\nIf you’ve ever wondered how data science, AI, robotics, and engineering actually come alive beyond textbooks — this is the place to find out.\nThrough these sessions, we go hands-on with the technologies shaping the present and future of industry, from data pipelines and AI automation to robotics innovation labs and professional growth frameworks. You’ll engage with curated content, collaborative exercises, and engineering-grade challenges that prepare you for serious, real-world projects — not superficial experience.\nFeatured Workshop Tracks # Professional Personal Branding 101\nYour zero-to-hero guide for Gen Z professionals entering the modern workforce. Understand how to brand yourself online and offline with authenticity and strategic purpose.\nData and AI Careers Under the Telescope\nExplore the breadth of data-driven roles — from data analysis and engineering to data science, machine learning, and artificial intelligence. Get insider insights into workflows, tools, and hiring expectations.\n","title":"Workshops \u0026 Camps","type":"workshops-camps"},{"content":" Your Zero-to-Hero Path to Industry-Ready Expertise # Explore comprehensive learning roadmaps for the world\u0026rsquo;s leading specialties on the Roadmaps page—from data analysis and data engineering to data science, machine learning, artificial intelligence, Mechatronics, and robotics.\nNow you\u0026rsquo;re on the learning path! We\u0026rsquo;ll guide you step-by-step from true zero to factory floors and cutting-edge research labs. Our roadmaps are intensive, demanding a strong believer to build rock-solid science foundations. They focus on practical knowledge with essential engineering and mathematical skills—no theoretical fluff or DIY YouTube hacks that amaze but never land real job offers.\nCross over to my Courses section for structured Python and MLOps training, and join Workshops and Camps \u0026hellip;\nReady to level up? Start with one of our courses today!\nFeatured Roadmaps Designed for You # Related pages to execute your roadmap # Courses for structured implementation. Workshops \u0026amp; Camps for live practical sessions. Mentorship Services for personalized guidance. Python Data Engineering \u0026amp; MLOps roadmap for data/AI path. Mechatronics Engineering roadmap for robotics/embedded path. ","externalUrl":null,"permalink":"/roadmaps/","section":"Your Zero-to-Hero Path to Industry-Ready Expertise","summary":"Your Zero-to-Hero Path to Industry-Ready Expertise # Explore comprehensive learning roadmaps for the world’s leading specialties on the Roadmaps page—from data analysis and data engineering to data science, machine learning, artificial intelligence, Mechatronics, and robotics.\nNow you’re on the learning path! We’ll guide you step-by-step from true zero to factory floors and cutting-edge research labs. Our roadmaps are intensive, demanding a strong believer to build rock-solid science foundations. They focus on practical knowledge with essential engineering and mathematical skills—no theoretical fluff or DIY YouTube hacks that amaze but never land real job offers.\nCross over to my Courses section for structured Python and MLOps training, and join Workshops and Camps …\nReady to level up? Start with one of our courses today!\nFeatured Roadmaps Designed for You # Related pages to execute your roadmap # Courses for structured implementation. Workshops \u0026 Camps for live practical sessions. Mentorship Services for personalized guidance. Python Data Engineering \u0026 MLOps roadmap for data/AI path. Mechatronics Engineering roadmap for robotics/embedded path. ","title":"Your Zero-to-Hero Path to Industry-Ready Expertise","type":"roadmaps"},{"content":" أعلن جاهزيتي التامة للتعاون الاحترافي واستضافة محاضرات أو ورش عمل في المدن السورية دون أي مقابل مادي. أقدّم هذا العرض لأن هدفي بعيد المدى، ورؤيتي للمستقبل أكبر بكثير من أي عائد مالي قريب.\nخلفية هذه الدعوة في مقالتي: بناء الإمبراطوريات لا يأتي بالصدفة.\nما الذي أقدّمه # محاضرة تفاعلية (ساعة إلى ساعتين). ورشة عمل تطبيقية (3 إلى 5 ساعات). في مجالات علوم البيانات وتعلّم الآلة، منهجية البحث العلمي والنشر الأكاديمي، الهوية الرقمية الاحترافية، والميكاترونكس والروبوتيك.\nملاحظة حول النطاق: هذه الدعوة مخصّصة للمحاضرات وورش العمل القصيرة فقط. أما الحقائب التدريبية المتكاملة فلها مسار تنسيقي وتعاقدي مختلف وهي مأجورة بطبيعتها — يمكنك الاطّلاع عليها في صفحة الإرشاد والاستشارات. شروط التعاون # حتى تكون الفعالية على المستوى الذي يليق بمحتواها وبجمهورها، يُشترط:\nبيئة احترافية ومجهّزة — قاعة مناسبة، جهاز عرض وشاشة، نظام صوتي، إنترنت سريع ومستقر، وتكييف/تدفئة ملائمة. تنظيم وإدارة احترافية — فريق منظّم يدير القاعة ودخول الحضور والالتزام بالجدول الزمني. تغطية إعلامية تليق بمستوى المحتوى المطروح — تصوير فوتوغرافي وفيديو، وتغطية عبر منصّات مناسبة. كيف تقدّم طلب التعاون # الطريقة الرسمية لتنسيق أي استضافة هي تعبئة النموذج أدناه. يرجى إدراج كافة التفاصيل واقتراح ثلاثة مواعيد على الأقل حتى نتمكّن من مقاطعتها مع جدول الأعمال.\nLoading… أسئلة أخرى أو تواصل مباشر # للأسئلة التي لا يغطيها النموذج، أو لإرسال ملف تعريف جهتكم (Portfolio / Company Profile)، راسلوني على: contact@mulhamfetna.com\nأعتذر سلفاً عن عدم الرد على أي عرض غير مكتمل، أو يشمل ما يتجاوز المحاضرات وورش العمل — فالحقائب التدريبية الكاملة لها مسارها الخاص.\n","externalUrl":null,"permalink":"/work-with-me/invite/","section":"Work With Me","summary":" أعلن جاهزيتي التامة للتعاون الاحترافي واستضافة محاضرات أو ورش عمل في المدن السورية دون أي مقابل مادي. أقدّم هذا العرض لأن هدفي بعيد المدى، ورؤيتي للمستقبل أكبر بكثير من أي عائد مالي قريب.\nخلفية هذه الدعوة في مقالتي: بناء الإمبراطوريات لا يأتي بالصدفة.\nما الذي أقدّمه # محاضرة تفاعلية (ساعة إلى ساعتين). ورشة عمل تطبيقية (3 إلى 5 ساعات). في مجالات علوم البيانات وتعلّم الآلة، منهجية البحث العلمي والنشر الأكاديمي، الهوية الرقمية الاحترافية، والميكاترونكس والروبوتيك.\nملاحظة حول النطاق: هذه الدعوة مخصّصة للمحاضرات وورش العمل القصيرة فقط. أما الحقائب التدريبية المتكاملة فلها مسار تنسيقي وتعاقدي مختلف وهي مأجورة بطبيعتها — يمكنك الاطّلاع عليها في صفحة الإرشاد والاستشارات. شروط التعاون # حتى تكون الفعالية على المستوى الذي يليق بمحتواها وبجمهورها، يُشترط:\nبيئة احترافية ومجهّزة — قاعة مناسبة، جهاز عرض وشاشة، نظام صوتي، إنترنت سريع ومستقر، وتكييف/تدفئة ملائمة. تنظيم وإدارة احترافية — فريق منظّم يدير القاعة ودخول الحضور والالتزام بالجدول الزمني. تغطية إعلامية تليق بمستوى المحتوى المطروح — تصوير فوتوغرافي وفيديو، وتغطية عبر منصّات مناسبة. كيف تقدّم طلب التعاون # الطريقة الرسمية لتنسيق أي استضافة هي تعبئة النموذج أدناه. يرجى إدراج كافة التفاصيل واقتراح ثلاثة مواعيد على الأقل حتى نتمكّن من مقاطعتها مع جدول الأعمال.\n","title":"استضافة ملهم | Invite Me to Speak","type":"work-with-me"}]