PeopleofSuperintelligence

Researcher

Peter Mostowsky

Also known as Петр Алексеевич Мостовский

Peter Mostowsky is a machine learning researcher who works on Gaussian processes, a model class for learning unknown functions while representing predictive uncertainty. With Viacheslav Borovitskiy, Alexander Terenin, Marc Deisenroth and others, they coauthored "Efficiently sampling functions from Gaussian process posteriors", which received an Outstanding Paper Honorable Mention at ICML 2020, and "Matérn Gaussian Processes on Riemannian Manifolds" for NeurIPS 2020. By 2021, they were a research engineer in the Department of Mathematics and Computer Science at St. Petersburg State University; the AISTATS paper "Matérn Gaussian Processes on Graphs" and the JMLR paper "Pathwise Conditioning of Gaussian Processes" also appeared in 2021. They headed the author list of the GeometricKernels package paper, posted on arXiv in July 2024 and published in JMLR in 2025, and a March 2026 revision of that paper gives Independent as their affiliation.

Timeline

  1. 2020

    Sampling Gaussian process posteriors

    With James Wilson, Viacheslav Borovitskiy, Alexander Terenin and Marc Deisenroth, they coauthor "Efficiently sampling functions from Gaussian process posteriors" for the 37th International Conference on Machine Learning in 2020. The paper separates the prior from the data to draw posterior samples at a fraction of the usual cost.

  2. 2020

    An ICML honorable mention

    At ICML 2020, the sampling paper receives an Outstanding Paper Honorable Mention.

  3. 2020

    Matérn kernels on manifolds

    Also in 2020, they coauthor "Matérn Gaussian Processes on Riemannian Manifolds" with Viacheslav Borovitskiy, Alexander Terenin and Marc Deisenroth, presented as a NeurIPS 2020 poster. The paper computes the kernels of Matérn Gaussian processes on compact Riemannian manifolds through spectral theory of the Laplace-Beltrami operator, which lets standard scalable techniques such as inducing point methods train them.

  4. By 2021

    Research engineer at St. Petersburg State

    By 2021, they work as a research engineer in the Department of Mathematics and Computer Science at St. Petersburg State University, with scientific interests in homological algebra and category theory.

  5. 2021

    Matérn processes on graphs

    In 2021, they coauthor "Matérn Gaussian Processes on Graphs" with Viacheslav Borovitskiy, Iskander Azangulov, Alexander Terenin, Marc Deisenroth and Nicolas Durrande for AISTATS, the 24th International Conference on Artificial Intelligence and Statistics. The paper builds on the stochastic partial differential equation characterization of Matérn Gaussian processes to study their analog on undirected graphs.

  6. 2021

    Pathwise conditioning in JMLR

    The Journal of Machine Learning Research publishes "Pathwise Conditioning of Gaussian Processes" in 2021, written with James T. Wilson, Viacheslav Borovitskiy, Alexander Terenin and Marc Peter Deisenroth. The paper derives a family of approximations for efficiently sampling Gaussian process posteriors and applies them to settings such as global optimization and reinforcement learning.

  7. 2024–2025

    The GeometricKernels package

    GeometricKernels is a Python package for heat and Matérn kernels on manifolds, meshes and graphs, and its companion paper appears on arXiv in July 2024 with them as first of nine authors. The Journal of Machine Learning Research publishes the paper in 2025.

  8. March 2026

    A 2026 GeometricKernels revision

    A second version of the GeometricKernels paper appears on arXiv in March 2026, with their affiliation given as Independent.

  1. Google had the most people shaping AI every year from 2014 to 2025Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.
  2. The United States employs an estimated 48% of people shaping AIWe estimate where the people shaping AI work today, how universities, companies and research institutes split across countries, and how country shares have changed since the 1960s.
  3. 43% of living people shaping AI work outside their country of birthWe estimate that 43% of living people shaping AI work outside their country of birth (520 of 1,198 people with known birth and employment countries).
  4. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionWe estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Peter Mostowsky yet. Send us a free-licensed photo.

  1. Efficiently sampling functions from Gaussian process posteriors proceedings.mlr.press. Checked 22 October 2026.
  2. ICML 2020 Awards icml.cc. Checked 22 October 2026.
  3. Matérn Gaussian Processes on Riemannian Manifolds neurips.cc. Checked 22 October 2026.
  4. SPbU | Department of Mathematics and Computer Science » Peter Mostowsky web.archive.org. Checked 22 October 2026.
  5. Matérn Gaussian Processes on Graphs proceedings.mlr.press. Checked 22 October 2026.
  6. Pathwise Conditioning of Gaussian Processes jmlr.org. Checked 22 October 2026.
  7. The GeometricKernels Package: Heat and Matérn Kernels for Geometric Learning on Manifolds, Meshes, and Graphs jmlr.org. Checked 22 October 2026.
  8. [2407.08086] The GeometricKernels Package: Heat and Matérn Kernels for Geometric Learning on Manifolds, Meshes, and Graphs arxiv.org. Checked 22 October 2026.
  9. The GeometricKernels Package: Heat and Matérn Kernels for Geometric Learning on Manifolds, Meshes, and Graphs (arXiv HTML v2) arxiv.org. Checked 22 October 2026.