I am now a Machine Learning Research Scientist at Xaira Therapeutics, based in the London office, working with Joseph Watson, Tor Fjelde, Francisco Vargas and Justin Barton. We are developing and training generative models for antibody design.

Before that I was a Postdoctoral Research Associate in the Cambridge Machine Learning Group, part of the Prosperity grant led by Prof Richard Turner and Prof José Miguel Hernández-Lobato, and in partnership with Microsoft Research. Previously I was an EPSRC Postdoctoral Research Associate at the University of Oxford affiliated with the Department of Statistics, and part of the Oxford Statistical Machine Learning Group, working with Pr Arnaud Doucet and Pr Yee Whye Teh. Previous to that, I was a DPhil candidate in the same research group, supervised by Pr Yee Whye Teh an Ryota Tomioka from Microsoft Research. Prior to that, I was a graduate enrolled in the Mathematiques, Vision, Learning (MVA) master at Ecole Normale Supérieure Paris-Saclay, with an emphasis on machine learning. I previously studied at Ecole des Ponts ParisTech in the mathematics and computer science department.

Interests

My research interests centre around `Generative AI’ for the Natural Sciences, with a focus on encoding problem symmetries and geometrical constraints. I am particularly interested in molecular science and weather science.

News

• February 2024. I’ve joined Xaira Therapeutics as a Machine Learning Research Scientist in London, working with Joseph Watsonon developing generative models for antibody design. Looking forward to making new therapeutics!
• January 2024. With my friends Tor and Vincent we wrote a short-ish blog post introducing flow matching, from a normalising flow perspective. Hope it’ll be useful!
• May-July 2023. Recent work on protein design with diffusion models I’m really excited about!
  • Accepted to ICML: SE(3) Diffusion Model with Application to Protein Backbone Generation,
  • Accepted to Nature: Broadly Applicable and Accurate Protein Design by Integrating Structure Prediction Networks and Diffusion Generative Models,
  • Accepted to TMLR: Diffusion Models for Constrained Domains and follow up work Metropolis Sampling for Constrained Diffusion Models.
• October 2022. Our work Riemannian Score-Based Generative Modeling where we extend score-based diffusion models to a broad class of manifolds has been accepted @ Neurips 2022 as an `oustanding paper’ :)
• June 2022. I’m joining the Cambridge Machine Learning Group as Postdoctoral Research Associate! Excited to work with Prof Richard Turner and Prof José Miguel Hernández-Lobato.
• Dec 2021. I passed my viva with no correction! I’d like to thank Arnaud Doucet and Max Welling for putting the effort to be my examinors.
• Sept 2021. Thesis submitted! Other great news, our work with Adam Foster On Contrastive Representations of Stochastic Processes has been accepted at NeurIPS 2021.
• August 2021. Starting a new postdoctoral position in the same great OxCSML research group, under the advisory of Pr Arnaud Doucet.
• Dec 2020. I’m co-organizing the first Differential Geometry meets Deep Learningworkshop, and we have our paper Riemannian Continuous Normalizing Flows accepted at the Conference on Neural Information Processing Systems (NeurIPS) 2020.
• July 2020. I passed my confirmation of status (last step before the PhD submission).
• Sept 2019. I’m starting an internship at FAIR New York with Max Nickel :)