AI Research Engineer / Data Scientist
| Skill | Description |
|---|---|
| 🐍 Python | 7+ years as my main language |
| 🔥 PyTorch | 6+ years as my main framework. Active on the forums, contributor on GitHub. |
| 🤖 Machine Learning | Modeling, optimization, pipelines, experimentation; sklearn, JAX enthusiast |
| 📊 Data Science | Analysis, feature engineering, visualization, statistical testing |
| 🦙 AI | Computer Vision, LLMs, LangChain, RAG, MCP, agents, model fine-tuning |
| 🚀 CI/CD | Testing, automatic documentation building, deployment, docker |
| 💻 C | The first language I ever learned. I made embedded ML algorithms in C at Withings |
| 🗄️ SQL | Data querying, analytics |
| 🌐 Open Source | Maintainer of TorchJD; Contributor of PyTorch, Torchmetrics, LibMTL |
- TorchJD: PyTorch library for Jacobian descent, enabling multi-objective neural network training, based on the theory from my paper Jacobian Descent For Multi-Objective Optimization and related research.
- RagThisCode Setup an MCP server to easily do Retrieval Augmented Generation of files from GitHub repos. Also comes with a chatbot web app using an agent connected to the MCP server, to chat "with" a repo.
- Federated Learning for Malware Detection in IoT Devices (Computer Networks, first author)
- 400+ citations, Best Paper Award
- Jacobian Descent for Multi-Objective Optimization (arXiv, co-first author)
- My best paper. Rejected from NeurIPS, ICLR and ICML but still keeping faith in it!
| Degree | Institution | Notes |
|---|---|---|
| 🧑🎓 MSc in Data Science | 🇨🇭 EPFL | Mathematics, algorithms, ML theory, deep learning, data analysis |
| 🧑🎓 BSc in Communication Systems | 🇨🇭 EPFL | Mathematics, physics, computer science, telecom |
| 🌏 Exchange Year (1 year) | 🇸🇬 NUS | Mathematics, computer science, telecom |
I'm looing for a full-time job in AI research and development. I'm also open to freelance missions on anything within my skillset.
- LinkedIn: valérian-rey