Research

I am currently focused on zero-knowledge verification for frontier AI training. Other research threads are mostly on hold or lower priority while this programme is active.

Current Focus

I work on technical verification primitives for frontier AI training: ways to make claims about large training runs checkable while preserving confidentiality. This line connects cryptographic proof systems, distributed training, hardware/network observations, and AI governance.

I am working closely with the GPAI Policy Lab verification programme and with the contributor group listed on that page, including Pierre Peigné, Ky Nguyen, Tom David, Emeric Bergmann, and Dimitris Kolonelos.

Preprints

Publications and Accepted Papers

Talks

Notes

Introduction au Domaine de Recherche

Théorie géométrique des modèles, 2021.

PDF

PhD manuscript

Variations on the group configuration, ENS-PSL, 2025.

HAL: tel-05419277

CV

Current CV, July 2026.

PDF