I am a postdoctoral researcher in the Theory of Machine Learning lab at EPFL, working with Nicolas Flammarion to deepen the theoretical understanding of Large Language Models and their advanced learning capabilities. My research interests include multi-armed bandits, reinforcement learning, online learning, imitation learning, optimization, statistics, and machine learning in general. I focus on developing theoretically well-founded and practically applicable algorithms.
I received my PhD in Computer Science from the University of Lille, where I worked under the direction of Emilie Kaufmann and Rémy Degenne in the Inria Scool team. My PhD research focused on pure exploration problems for stochastic multi-armed bandits. My thesis endeavored to establish the Top Two approach as a principled methodology offering nearly optimal theoretical guarantees alongside state-of-the-art empirical performance. My thesis touched upon various aspects of bandit theory: parametric and non-parametric classes of distributions, structural assumptions on their means, and different identification problems. During my PhD, I had the opportunity to visit Nicolò Cesa-Bianchi at the University of Milan in the Laboratory for Artificial Intelligence and Learning Algorithms for three months, where I worked on adversarial regret minimization for contextual linear bandits.
Before my PhD, I graduated from Ecole Polytechnique and ETH Zurich. I conducted my Master’s thesis in the Learning & Adaptive Systems group of Andreas Krause, where I studied pure exploration for combinatorial bandits with semi-bandit feedback.
PhD in Computer Science, 2021-2024
Scool (Inria) / CRIStAL (CNRS) / Univ. Lille
MSc ETH in Data Science, 2018-2020
ETH Zurich
Diplôme d'Ingénieur (MSc), 2015-2018
École Polytechnique