|
|
Hey, I’m Youcef — an AI/ML Engineer and Full Stack Developer focused on building intelligent decision-support systems, RAG document pipelines, Computer Vision, MLOps & CI/CD workflows, and scalable Web/Mobile Applications.
I build projects that start as “this idea is insane” and somehow become “wait, this actually works.”
The model hallucinates. The backend panics. The frontend shifts by 2 pixels for no reason. The database silently judges everyone.
And yet... we ship. 🚀
| Class | Role | Passive Ability | Ultimate |
|---|---|---|---|
| 🧙 RAG Wizard | Document intelligence | Extracts truth from PDFs & vector stores | Citation Storm |
| 🤖 Agentic AI Summoner | Tool-using AI workflows | Automates repetitive cognitive loops | Multi-Agent Swarm |
| 👁️ Computer Vision Hunter | Pixel analysis & defect detection | Spots anomalies in the wild | Frame-by-Frame Judgment |
| 🛡️ Full Stack Tank | Product engineering | Carries frontend + backend + database | Deploy Without Fear |
| 🧪 MLOps Alchemist | Pipeline automation & drift tracking | Turns raw data into reproducible models | Continuous Retraining Ritual |
mindmap
root((Youcef))
AI & ML
PyTorch
Scikit-learn
TensorFlow
Pandas & NumPy
Model Evaluation
RAG & GenAI
LangChain
Chunking & Embeddings
Vector Search Qdrant
Reranking
Citation Grounding
MLOps
MLflow
DVC
Docker
CI/CD Pipelines
Data Drift Monitoring
Computer Vision
OpenCV
Image Classification
Anomaly Detection
Object Detection
Full Stack
React
Next.js
TypeScript
FastAPI
Django REST
PostgreSQL
Infra & Cloud
Docker
GitHub Actions
Vercel
Linux
Git
pie title Current Power Allocation
"AI / ML / RAG" : 35
"Full Stack & APIs" : 25
"MLOps & Pipelines" : 15
"Computer Vision" : 15
"Debugging at 2 AM" : 10
| Quest | Status |
|---|---|
| Build full-stack AI applications | ✅ Completed |
| Architect end-to-end RAG pipelines | ✅ Completed |
| Build automated MLOps & CI/CD workflows | ✅ Completed |
| Deploy intelligent agentic systems | 🟩 Active |
| Train computer vision anomaly detectors | 🟩 Active |
| Optimize low-latency vector retrieval | 🟨 In Progress |
| Defeat flaky dependencies & CUDA mismatch | ♾️ Eternal Quest |
| Touch grass | ❌ Optional DLC |
🟢 Level 1: API returns 500. Choose your move.
A) Refresh repeatedly and hope
B) Blame the frontend
C) Check logs like a responsible engineer
✅ Correct: C
Reward: +10 sanity, +1 backend mastery, 500 error took emotional damage
🟡 Level 2: RAG output is hallucinating. Choose your spell.
A) Increase temperature because chaos is art
B) Improve chunking, retrieval, reranking, and context grounding
C) Tell users "it's a feature"
✅ Correct: B
Reward: +20 truthfulness, +15 retrieval accuracy, hallucination debuffed
🔴 Level 3: Production broke at 2 AM. Final boss music starts.
Boss: NullPointerDragon
HP: 9999
Weakness: stack traces
Move sequence: logs → reproduce → root cause → patch → test → deploy → pretend calm
🏆 Victory unlocked: Production Survivor
🟣 Secret Level: The model works locally but fails in deployment.
Possible causes:
- Missing environment variable
- Wrong model path or CUDA driver mismatch
- Container memory limit
- Dependency mismatch between torch and torchvision
- The cloud simply woke up and chose violence
Reward: +30 DevOps XP, +1 thousand-yard stare
| Achievement | Description |
|---|---|
| 🧙 RAG Wizard | Built retrieval-augmented generation systems |
| ⚔️ API Knight | Survived high-concurrency REST API battles |
| 🐉 Bug Slayer | Defeated errors that had no business existing |
| 🧠 Tensor Tamer | Trained and fine-tuned models without losing sanity |
| 🏗️ Full Stack Builder | Built apps from database schema to modern UI |
| 🚀 Ship Happens | Deployed projects despite the universe resisting |
| 🕵️ Log Detective | Found bugs through terminal archaeology |
| 🔥 Overengineering Enjoyer | Made the architecture 3x cooler than required |
