« I don't replace humans. I give them superpowers. »
(Also: I once treated a lion. The lion did not submit a PR. I learned humility.)
Most people pivot from vet school → clinic. classAbdelAtia:
degrees= ["PhD (Biopharmaceutics)", "DMV (Veterinary Medicine)"]
superpower="Speaks cells, Python, and LLMs — often in the same sentence."origin_story="Fell in love with code as a kid. Never recovered."side_quests= {"countries": 64, "certifications": 8, "papers_per_year": "300+"}
defdaily_loop(self):
return"☕ → 🐛 → 🔀 merge → 📚 read → repeat"deftreat(self, patient):
ifpatient.species=="Panthera leo":
return"Handled. No unit tests required."returnself.rag_pipeline.ingest(patient.data) |  
📖 the long version (still readable)🧒 Origin — Self-taught coder since childhood. BASIC → C → Python → agents. 🎓 Academic — DMV (yes, lions in Africa). PhD in biopharmaceutical sciences. 🚀 Now — RAG & agentic AI architect. IBM + Harvard certified (2025–2026). Common thread: biology gave me rigor; code gave me leverage; AI lets me fuse both without asking permission. |
Where I leave code that outlives my coffee.
OMML→MathML stack (radicals, n-ary, eqArr, accents…) shipped in stable release notes — 2026-04-17
⭐ superdoc-dev/superdoc Built the OMML → MathML layer: equation arrays, matrices, n-ary operators, group chars, phantoms, accents, pre-sub-superscripts, box/borderBox. Word math renders natively in the browser. #2754 · #2753 · #2752 · #2751 · #2750 · #2749 · #2748 · #2747 | ⭐ HKUDS/RAG-Anything Multimodal RAG ingestion: Docling CLI → Python API (#261), OMML equation extraction for DOCX (#262), public asset URL docs (#275). |
📦 full merge log — 25 upstream repos (+ 6 on Abdeltoto repos = 31 total)
| Project | ★ | PRs | What shipped |
|---|---|---|---|
| RAG-Anything |  | 3 | Docling API, OMML extract, asset URL docs |
| Awesome-Bioinformatics |  | 1 | Added Scanpy, EvoDiff |
| Apache Hamilton |  | 2 | Async builder typing + docs (#1559, #1560) |
| awesome-neuroscience |  | 1 | DeepLabCut, CaImAn, Elephant |
| eLabFTW |  | 1 | ELN account-creation security (#6786) |
| SuperDoc |  | 8 | Full OMML→MathML conversion stack |
| PyLabRobot |  | 1 | Thermocycler hold-time error (#1010) |
| awesome-computational-biology |  | 1 | Curated comp-bio tools |
| gsd-for-cursor |  | 1 | Fixed hooks path ~/.cursor/ (#3) |
| DeepCritical |  | 1 | MAFFT MCP server (#231) |
| SCHNAPPs |  | 1 | Fix scExOrg in sample data (#8) |
| q2-dada2 |  | 1 | n_reads_learn help text (#191) |
| palantir-mcp |  | 1 | Windows shell: true fix (#25) |
| ProBioPred |  | 1 | Modern pyproject.toml scaffolding (#1) |
| vllm-omni_AA |  | 1 | L5 stability test framework (#1) |
| SENAITE |  | 1 | Docker quick-start docs (#2) |
Recent upstream merges:SCHNAPPs #8
Open upstream PRs:frictionless-py #1775 / #1776
Closed without merge (reopenable):airflow #67933
Totals from author:Abdeltoto is:pr is:merged · ★ counts live from GitHub — Jul 2026
🔍 code reviews — 13 PRs on third-party OSS
scikit-learn/scikit-learn
— 2 reviews
pandas-dev/pandas
— 1 review
- #65738 — DOC: fix outdated matplotlib link (approved)
apache/airflow
— 1 review
- #67905 — Trigger DAG params grid layout
| 🔍 RAG Systems Docs → knowledge → action Hybrid search · rerank · citations | 🤖 AI Agents Multi-agent orchestration ReAct · MCP · workflows | 🧬 Bioinformatics FASTQ → species ID 16S · BLAST · metagenomics | ⚡ Production APIs FastAPI · Docker · K8s Monitored · scalable |
🤖 IBM RAG & Agentic AI — 8-course specialization
LangChain · LangGraph · CrewAI · AutoGen · MCP · Vector stores · Prompt engineering · Multi-agent systems
🎓 Harvard CS50 — 7 courses completed
CS50x ✅ CS50P ✅ CS50 SQL ✅ CS50AI ✅ CS50W ✅ CS50R ✅ CS50 Cybersecurity ✅
🤖 AI Frameworks & Agents
🧠 LLMs · RAG · Vector DBs
🧬 Bioinformatics & Lab Tech
☁️ DevOps · Cloud · Data
🔍 Enterprise RAG Platform Stack: LangChain · LangGraph · Pinecone · Cohere · GPT-4.1Features: multi-source ingest · hybrid search · self-correcting RAG · citationsImpact: "90% faster research" | 🤖 Multi-Agent Research System Stack: LangGraph · CrewAI · Claude · GPT-4.1 · MCPAgents: Researcher · Analyst · Writer · CriticImpact: "20h/week automated" |
🧬 SeqDNA 16S Stack: Flask SPA · BLAST+ · SILVA 138.2 · PERMANOVA · BIOM exportFeatures: 16S pipeline · PDF reports · SaaS tiersImpact: "FASTQ → species ID in minutes" | ⚡ Intelligent Pipelines Stack: Airflow · Docker · PostgreSQL · RedisFeatures: self-healing · anomaly detection · auto-scalingImpact: "Zero manual intervention" |
classDevPhilosophy:
RULES= [
"If you do it twice → automate it",
"Clarity beats cleverness. Always.",
"Ship fast, iterate faster",
"Good code is code you can delete",
"Best prompt = the one you never write twice",
]
MOTTO="In the name of the Prompt, the Model, and the Holy Iteration 🙏" | - ⚠️ CaffeineDependency: unstable below 3 cups- ⚠️ TabOverflow: 50+ tabs is "normal"- ⚠️ RefactorLoop: "quick fix" → 5h rewrite- ⚠️ CertificationAddiction: 8 and counting+ ✅ RAGObsession: building retrieval 24/7 (WAI)+ ✅ AgentAddiction: agents for everything (feature)+ ✅ HarvardEffect: explains life with O(n) (feature)+ ✅ LionExperience: humility installed permanently |
