I build machine learning systems across transformers, neural time-series modeling, and performance-oriented engineering.
My recent work focuses on seizure onset zone localization from intracranial EEG (SEEG), with emphasis on transformer architectures, contrastive representation learning, and cross-patient generalization.
I am interested in applied ML roles where research taste matters, but where the final output still has to be engineered, tested, shipped, and explained clearly.
I grew up in a small town in France, trained as an electrical and computer engineer, and completed my M.S. in Electrical and Computer Engineering at Georgia Tech.
Research, systems, and tooling
Neural decoding framework for seizure onset zone localization, using transformer-based sequence modeling and contrastive pre-training for cross-patient generalization.
Market-making research platform built on reconstructed CME Globex WTI Crude Oil Level-3 order-book data, featuring FIFO queue modeling, latency-aware paper trading, interactive replay tooling, and strategy evaluation under realistic microstructure conditions.
GPU ray tracer with coordinator/worker execution, CUDA rendering kernels, UDP-based task distribution, and interactive SFML visualization.
Production newsletter pipeline converting structured Markdown into styled, email-ready HTML for HKN Beta Mu chapter communications.