Boris Bonev

NVIDIA Research. Zurich, Switzerland

boris.jpg

Boris Bonev is a Senior Research Scientist at NVIDIA Research in Zurich. His research lies at the intersection of numerical methods and machine learning for science and engineering, with emphasis on reliable AI for weather and climate prediction.

He completed a Ph.D. in Applied Mathematics at the Swiss Federal Institute of Technology (EPFL) with Jan S. Hesthaven, after degrees in Physics and Aerospace Engineering from the University of Stuttgart. At NVIDIA, he worked under Anima Anandkumar and Alex Keller, and collaborated closely with Thorsten Kurth, Michael Pritchard, and William D. Collins. His recent work has focused on geometric neural operators for weather and climate prediction—including Spherical Fourier Neural Operators and related methods that have been adopted in FourCastNet, Huge Ensembles, ACE, and many other models. Earlier work spans hierarchical-matrix preconditioners and discontinuous Galerkin methods for wave problems.

Further detail is on the research and software pages.

Selected Publications

  1. FourCastNet 3: A geometric approach to probabilistic machine-learning weather forecasting at scale
    Boris Bonev, Thorsten Kurth, Ankur Mahesh, and 7 more authors
    2025
    arXiv preprint arXiv:2507.12144
    arXiv HTMLPDF
  1. GMD
    Huge ensembles–Part 1: Design of ensemble weather forecasts using spherical Fourier neural operators
    Ankur Mahesh, William D. Collins, Boris Bonev, and 5 more authors
    Geoscientific Model Development, 2025
  1. NeurIPS
    Attention on the Sphere
    Boris Bonev, Max Rietmann, Andrea Paris, and 2 more authors
    In Advances in Neural Information Processing Systems, 2025
  1. local_no.png
    Neural Operators with Localized Integral and Differential Kernels
    Miguel Liu-Schiaffini, Julius Berner, Boris Bonev, and 3 more authors
    In International Conference on Machine Learning, 2024
    HTMLPDF
  1. ace-preview.png
    ACE: A fast, skillful learned global atmospheric model for climate prediction
    Oliver Watt-Meyer, Gideon Dresdner, Jeremy McGibbon, and 5 more authors
    2023
    arXiv preprint arXiv:2310.02074
    arXiv HTMLPDF
  1. sfno_showcase.png
    Spherical Fourier Neural Operators: Learning Stable Dynamics on the Sphere
    Boris Bonev, Thorsten Kurth, Christian Hundt, and 4 more authors
    In International Conference on Machine Learning, 2023
    HTMLPDF Code