Lucas Meyer
I am currently a final year PhD student in computer science and applied mathematics at Université Grenoble Alpes in partnership with INRIA and EDF. My research, under the supervision of Bruno Raffin and Alejandro Ribes, focuses on training deep learning models for numerical simulation at scale.
Many engineering and scientific applications require faithful numerical simulations of partial differential equations. These simulations rely on traditional solvers that incur time and memory-intensive computation. Deep learning can alleviate this computational burden. However, supervised approaches need many runs of the same solvers whose slowness and memory footprint motivated the deep learning approach in first instance. I am interesting in training framework that leverage high performance computing resources to efficiently train deep learning on numerical simulations.
Before my PhD, I worked 2 years in the space industry, including at the European Space Agency, developing software and deep learning methods for remote sensing. In 2018, I received a M.Sc in computer science from Université de Montréal, working on optimization software for swarm of robots under the supervision of Giovanni Beltrame. I graduated in software engineering and applied mathematics from École Nationale des Ponts et Chaussées.