Quick Introduction

I am a postdoctoral researcher at DAVID lab, Université Versailles Saint-Quentin (UVSQ), working on structure elucidation for molecules and crystaline solids with Jérémie Cabessa, Thibault Charpentier, Marie-Pierre Gaigeot, and Mihai-Cosmin Marinica.

Prior to that, I conducted my PhD in the MATHERIALS team, INRIA Paris, and CERMICS lab, École Nationale des Ponts et Chaussées (ENPC), under the supervision of Éric Cancès and Antoine Levitt.

I thrive at the intersection of applied mathematics, computational physics, and machine learning. Each of my projects has involved learning a new field, working closely with researchers and developers to formalize open problems, and delivering concrete proof-of-concept implementations — from mathematical derivation to working code.

Research Interests

My core background is in the numerical analysis of electronic structure methods (quantum chemistry, condensed matter physics). Over the past year I have been working on deep learning for molecular science: I built from scratch a PyTorch framework benchmarking diffusion models and stochastic interpolants — using Transformer architectures — for molecular structure elucidation (code to be released in 2026).

More details in the Research and Code sections.

Keywords: deep learning (inverse design, diffusion and post-diffusion models, LLMs), electronic structure (wave function methods, open-shell systems, plane wave DFT), numerical optimization (numerical analysis, Riemannian optimization).

Email: laurent(dot)vidal(at)uvsq(dot)fr.