Ryan-Rhys Griffiths

Ryan-Rhys Griffiths

Welcome to my personal website and thank you for taking the time to visit!

Overview

I’m a Research Scientist working on AI in San Francisco. My current focus is on large reasoning models and in particular their capabilities in domains such as mathematics, chemistry, and chess. I serve as Director of Machine Learning and Partnerships for the Fields Model Initiative.

Other topics that interest me include probabilistic machine learning and AI4Science. I have worked on these topics at Meta, Huawei Noah’s Ark Lab, MILA, and the University of Cambridge, where I did my PhD in the Theory of Condensed Matter Group at the Cavendish Laboratory. Prior to that I did my undergrad in chemical physics at Imperial College London.

For a more granular picture of my research interests see my Google Scholar page.

Miscellaneous Facts

Research Collaboration

I receive a huge volume of emails and LinkedIn messages each day so please accept my apologies if I can’t reply to yours! Nonetheless, I have bandwidth for a small number of research collaborations. If you think I may be a useful supervisor for your project please see the instructions in contact!

News

Oct 20, 2026 I will serve as an Area Chair for ICML 2027.
Aug 28, 2026 Our paper Large language models as uncertainty-calibrated optimizers for experimental discovery has been published in Nature Machine Intelligence.
May 05, 2026 I co-organized the Optimisation and Post-Bayesian Inference in Machine Learning (OPTIMAL) workshop at AISTATS 2026 in Tangier, Morocco.
Oct 17, 2025 I received a top reviewer award (top 8%) for NeurIPS 2025.
Aug 22, 2025 I will serve as a reviewer for AISTATS 2026.

Selected Publications

  1. chatgpt.png
    Mathematical Capabilities of ChatGPT
    Simon Frieder, Luca Pinchetti, Ryan-Rhys Griffiths, and 4 more authors
    Advances in Neural Information Processing Systems, 2024
  2. reasoning.png
    Training a Scientific Reasoning Model for Chemistry
    Siddharth M Narayanan, James D Braza, Ryan-Rhys Griffiths, and 7 more authors
    Advances in Neural Information Processing Systems, 2025
  3. aviary.png
    Aviary: training language agents on challenging scientific tasks
    Siddharth Narayanan, James D Braza, Ryan-Rhys Griffiths, and 8 more authors
    arXiv preprint arXiv:2412.21154, 2024
  4. gauche.png
    GAUCHE: a Library for Gaussian Processes in Chemistry
    Ryan-Rhys Griffiths, Leo Klarner, Henry Moss, and 8 more authors
    Advances in Neural Information Processing Systems, 2024
  5. phd.png
    Gaussian Processes at Extreme Lengthscales: From Molecules to Black Holes
    Ryan-Rhys Griffiths
    University of Cambridge, 2023
  6. hebo.png
    HEBO: Pushing the Limits of Sample-Efficient Hyper-Parameter Optimisation
    Alexander I Cowen-Rivers, Wenlong Lyu, Rasul Tutunov, and 8 more authors
    Journal of Artificial Intelligence Research, 2022
  7. copilot.png
    Data for Mathematical Copilots: Better Ways of Presenting Proofs for Machine Learning
    Simon Frieder, Jonas Bayer, Katherine M Collins, and 8 more authors
    arXiv preprint arXiv:2412.15184, 2024