I build ML products end to end, with a focus on bio/health R&D workflows and production-ready deployment. Recent work includes protein-focused ML pipelines and genomics-driven modeling. RL and 3D simulation are personal interests; I also ship interactive web experiences with Next.js and Three.js. Pragmatic deployment: Docker on Ubuntu, CI/CD, and reproducible pipelines.
- MISTRAL-AI-MCP-HACKATHON: MCP-first Clash Royale demo for Mistral AI's MCP Hackathon, with a Next.js UI and an MCP agent playing in real time.
- Stealth bio/health ML product work (protein-focused ML pipelines, genomics-driven modeling, production deployment).
- RL-in-3D-Physics-Engine: Python rigid-body physics + RL sandbox with a quadruped agent.
- Poker-GTO: CFR+ solver with a PyTorch policy approximation and a Next.js UI.
- Physics-Engine-in-C: C-based multi-pendulum simulations with CSV outputs and Python visualization.
- ISIC-2024-Hackathon: Skin lesion classification using CNNs and gradient boosting.
- Fighting-Tanks-Reinforcement-Learning: Multi-agent RL tank battle environment.
- ML productization, evaluation, and deployment (bio/health)
- 3D simulation and physics
- Real-time web apps (Next.js, React, Three.js)
- Email: eliott.valette@hec.edu
- LinkedIn: https://www.linkedin.com/in/eliott-valette/
- Kaggle: https://www.kaggle.com/eliottvalette
- LeetCode: https://leetcode.com/eliottvalette/
Last updated: February 2026


