CV
Education
CentraleSupélec 2023–2026
Engineering Degree (Diplôme d’Ingénieur), Mathematics and Data Science
Université Paris-Saclay 2025–2026
Master 2 (M2), Mathematics and Artificial Intelligence
Mention Bien (“High Honors”), 14.76/20
Beihang University 2025
Exchange semester, AI for engineering
La Martinière Monplaisir 2021–2023
Preparatory classes (CPGE), mathematics and physics, Bachelor’s level
Experience
National University of Singapore, LEMA group (Yuejiang Liu) From Jan. 2027
Incoming · six-month research position in robot learning
MBDA Sep. 2025 – 17 Sep. 2026
Generative AI Engineer · Apprenticeship
Multi-agent systems built around large language models.
TotalEnergies Digital Factory Nov. 2023 – Aug. 2025
Data Scientist · Apprenticeship
Models for electric-vehicle charging: pricing, and which stations are open to the public.
Selected research and projects
- Few-Step Field Regularity: exact Gaussian analysis of probability paths, field regularity, and finite-step numerical error. Sole author; accepted as a poster at the NeurIPS 2026 GDDL workshop (non-archival).
- Ego-World JEPA: action-conditioned latent world models and MPPI planning on PushT, with a documented negative result.
- Adaptive Latent Search: learned and heuristic stopping rules for compute allocation during latent optimization.
- Decision Transformer for LOB trading: five-seed offline-RL benchmark on FI-2010, co-developed with Côme Genet.
- Proximal Diffusion Models: PyTorch reproduction of published proximal diffusion methods.
See Research for questions, results, limitations, and public artifacts.
Technical focus
Generative modeling; flow matching and diffusion; latent world models; reinforcement learning and planning; numerical evaluation; reproducible experimentation in Python and PyTorch.
Service, teaching, and leadership
- Reviewer, NeurIPS 2026 GDDL workshop.
- Practical AI workshops with Automatants, CentraleSupélec’s AI association (2024–2025).
- Led a six-person team running network infrastructure for 450+ members as president of Rézo Metz-Rennes Fédérés.
Languages
French (native) · English (professional)