PhD in Biophysics · Computational Biology & Data Science · Machine Learning
I am a postdoctoral researcher in biophysics and AI, with a PhD in Physics and experience in computational modeling, biomedical signal processing, and algorithmic trading. I am now transitioning into quantitative research and systematic trading.
During my PhD in biophysics at the University of Bordeaux, I studied how mechanical forces shape stem-cell differentiation using microfluidic systems, advanced microscopy and mathematical modeling. This work required high-dimensional data analysis, statistical inference and robust simulations.
I then completed a certified AI and Data Science program at CentraleSupélec & OpenClassrooms, and I am currently a postdoctoral researcher at IHU Liryc, where I develop machine-learning-based ECG imaging methods and deep learning algorithms for patient-specific electrode localization.
Since 2022, I have been actively studying financial markets and designing algorithmic trading strategies, experimenting with reinforcement learning, FinRL frameworks and LLM-guided information processing. I now seek to bring this experience to a quantitative research team.
Backtest example of a systematic equity strategy with technical signals, trade points, equity curve and normalized comparison. SPY with 20× leverage, traded weekly.
Design and backtesting of equity strategies using FinRL-style reinforcement learning frameworks combined with LLM-based information filtering for risk-aware decision-making. Deployed on a virtual portfolio since 2022 with consistently positive performance.
ML-driven ECG imaging pipelines reconstructing epicardial potentials and localizing arrhythmogenic substrates from non-invasive measurements, with deep learning models for patient-specific electrode localization.
Deep learning pipelines for image classification and biomarker prediction, including training, evaluation and deployment-ready code.
End-to-end tag recommendation system for textual data, exposed via a FastAPI backend that serves real-time tag suggestions.