Florentin Coeurdoux
Research scientist and quantitative researcher at CFM. Working at the intersection of generative modeling, stochastic interpolants, and production machine learning.
About
I am a VP Research Scientist at CFM (Capital Fund Management), a systematic hedge fund in Paris. I lead a quantitative research team building and deploying machine learning models for live investment strategies.
In parallel, I conduct foundational research on generative models as part of the CFM AI Lab, in close collaboration with Eric Vanden-Eijnden (Courant / NYU), Stéphane Mallat (ENS / Flatiron Institute), and Jean-Philippe Bouchaud (CFM). I supervise research interns and PhD students.
I hold a PhD in statistics (2023), supervised by Nicolas Dobigeon and Pierre Chainais. During my doctorate, I was a visiting scholar at the Department of Statistics, University of Oxford, working with Arnaud Doucet.
Research Interests
Generative Models
Stochastic interpolants, flow matching, normalizing flows, and diffusion models for density estimation, sampling, and likelihood-free inference.
Bayesian Inference
Plug-and-play methods that embed deep generative priors into posterior sampling, enabling scalable inference in high-dimensional inverse problems.
Optimal Transport
Computational and geometric tools for comparing probability distributions, including sliced-Wasserstein distances and flow-based OT maps.
Works
2025–2026
Earlier Work
Service
- Reviewer for NeurIPS, ICML, ICLR, AISTATS