Louis Béthune
Research Engineer · Apple Machine Learning Research
Hi, I’m Louis. I’m a research engineer at Apple Machine Learning Research in Paris, where I train large generative models and build the codebase that runs them.
I have always loved working at the intersection of engineering, mathematical modeling, physics, and neuroscience. Deep-learning research is a nice place where science and multi-disciplinary engineering collide.
Nowadays, my attention goes to three things.
- Generative models. I work on multimodal generation across text, images, and audio, especially with masked discrete models.
- ML systems. I like large-model training, distributed computation, and I’ve kept a love of low-level programming from my early days.
- Biologically inspired methods. I am curious about predictive coding, back-propagation-free training, massively distributed asynchronous systems, and the expressiveness of dynamical systems.
News
- October 2026: The Design Space of Tri-Modal Masked Diffusion Models received a spotlight at NeurIPS 2026 and will be presented in Sydney
- August 2026: submissions to the DiffuLM workshop are open
- June 2026: Learning Unmasking Policies for Diffusion Language Models was selected for an oral at ICML 2026
- June 2024: 1-Lipschitz Neural Distance Fields received the Best Paper Award at the Symposium on Geometry Processing 2024