Projects

Ongoing:

• DAIMOS — Distributed AI Model training Optimization at Scale, ANR, 09/2026–08/2030. Coordinator of the work package on GNN training algorithms (approx. 140K EUR of 900K EUR total).
• HAISLED — Hybridizing AI and Large-scale Simulations for Engineering Design, seed funding for an international chair at ANITI, 07/2024–06/2028. PI (approx. 230K EUR).

Completed:

• STAML — Scalable Training Algorithms for Machine Learning, Toulouse-INP, call Emergences, Transversalités et International (ETI), 01/2025–12/2025. PI (approx. 15K EUR).
• Hybridization of Nonlinear Multiscale Solution Strategies for Large-Scale Multiphysics Applications — SNSF, 07/2024–06/2025. PI (approx. 123K CHF; forfeited due to joining Toulouse-INP).
• Multilevel training of DeepONets – multiscale and multiphysics applications — SNSF Postdoc.Mobility grant, hosted at Brown University, 06/2022–05/2024. PI (approx. 115K CHF).

Computational resources

• Jean Zay (GENCI) — Parallel training algorithms for scientific machine learning, 80,000 GPU node-hours (8 GPUs per node) on the French national supercomputer, 2024–2026 (approx. 76K EUR).