MSc Data Science · École Polytechnique & Télécom Paris · ML Engineer Intern @ Wiremind
- 🎓 MSc Data Science (M2DS) at École Polytechnique + Master of Engineering at Télécom Paris
- 💼 ML Engineer Intern @ Wiremind — working on time series forecasting
- 🧠 Deep interest in NLP — implemented a GPT-2 style transformer from scratch and fine-tuned it for email fraud detection
- ⚽ Built a real-time football RAG system — streams live match events via Kafka and lets you chat with an AI that knows exactly what's happening in the game
- 🔎 Built an AI pipeline for open-source license detection — combines ScanCode static analysis with Mistral AI to infer SPDX licenses, achieving 0.84 Repo-F1 on a benchmark of 100 GitHub repositories
- 🏆 Top 10 / 500+ in Morocco's National Competitive Programming Contest (2021)
- 🌍 French · English (C1+) · Arabic · Spanish
ML & Data
MLOps & Infrastructure
Other languages & tools
| Project | Description | Stack |
|---|---|---|
| gpt-from-scratch | GPT-2 transformer built from scratch — multi-head attention, causal masking, full training pipeline | Python · PyTorch · FastAPI · Docker |
| Real-time-football-rag | Streams live match events via Kafka, tracks the last 30 events as context, and powers an AI chatbot that answers questions about the game in real time | Python · Kafka · Mistral AI · TypeScript |
| capstone-project | Modular pipeline that downloads license files, scans them with ScanCode, and uses Mistral AI to infer SPDX licenses — 0.84 Repo-F1 on 100 repos | Python · Mistral AI |
| NBA | NBA analytics and visualisation | JavaScript · Python · Jupyter |