PhD thesis focused on the application of statistical physics methods to provide a theoretical understanding of deep neural networks and classical machine learning models and algorithms; Developed six carefully designed linear unit tests for evaluating out-of-distribution generalization capabilities during the internship at Facebook AI Research, and released the code on GitHub.
Research Experience
Lead Research Scientist at Jasper AI (since 2024), leading the Image Research team; Research Scientist at Stability AI (2023); First employee and Lead Research Scientist at Clipdrop (2021-2023), developing state-of-the-art computer vision models; Interned at Facebook Artificial Intelligence Research (2020), collaborating with Léon Bottou and David Lopez-Paz on robust machine learning.
Education
Ph.D. in Statistical Physics & Machine Learning Theory from Institut de Physique Théorique (IPhT - CEA Saclay, Université Paris-Saclay), supervised by Lenka Zdeborová and Florent Krzakala; Master of Science in Theoretical Physics from Ecole Normale Supérieure de Paris; Master of Science and Engineering from Ecole polytechnique; Bachelor of Science from Ecole polytechnique; attended Classe Préparatoire aux Grandes Ecoles at Lycée Henri IV in Paris.
Background
Research interests include statistical physics of disordered systems, information theory, statistical inference, and message-passing algorithms with applications to theoretical machine learning. Currently leading the Image Research team at Jasper AI, focusing on generative AI tasks.