PeopleofSuperintelligence

Researcher

Yuri Chervonyi

Also known as Iurii Chervonyi

Yuri Chervonyi is a researcher who moved from string theory to AI, and helped build AlphaGeometry 2 for Olympiad geometry at Google DeepMind. He earned a physics PhD at the University at Albany in 2017, with Oleg Lunin as dissertation chair, and later coauthored DeepMind papers on reinforcement learning for commercial cooling systems in 2022 and tokamak magnetic control in 2023. In 2024, he co-led the development of AlphaGeometry 2, part of the system that reached the silver-medal standard at the International Mathematical Olympiad, and in 2025 he was first author of its journal paper. In 2026, he was among the authors of the Aletheia math research agent paper. He works at Google DeepMind.

Badges

Timeline

  1. 2015

    Killing–Yano tensors in string theory

    In 2015, he and Oleg Lunin, both in the Department of Physics at the University at Albany, publish "Killing(-Yano) Tensors in String Theory". The paper constructs these tensors for a large class of charged black holes in higher dimensions and works out how they behave under string dualities.

  2. 2017

    A physics PhD at Albany

    He earns a PhD from the Department of Physics at the University at Albany in 2017, with Oleg Lunin as dissertation chair. His dissertation, "Hidden symmetries in string theory", extends hidden symmetries known from general relativity to stringy backgrounds in higher dimensions and constructs new integrable string theories.

  3. 2022

    Cooling buildings with reinforcement learning

    In 2022, he is a DeepMind coauthor of "Controlling Commercial Cooling Systems Using Reinforcement Learning", which reports live experiments at two real-world facilities with Trane Technologies. The reinforcement-learning system saves approximately 9% and 13% of energy at the two sites.

  4. July 2023

    Reinforcement learning for tokamaks

    In July 2023, he is among the authors of "Towards practical reinforcement learning for tokamak magnetic control". The paper reports up to 65% better shape accuracy in simulation and validates the upgraded controllers in experiments on the TCV tokamak.

  5. 2024

    AlphaGeometry 2 at the IMO

    He and Trieu Trinh lead the development of AlphaGeometry 2, which with AlphaProof forms the system that reaches the silver-medal standard at the 2024 International Mathematical Olympiad. The system scores 28 of 42 points, equivalent to the top end of the silver-medal category.

  6. 2025

    The AlphaGeometry2 paper

    In 2025, he is first author of "Gold-medalist Performance in Solving Olympiad Geometry with AlphaGeometry2" in JMLR, which reports an overall solving rate of 84% on geometry problems from the last 25 years, up from 54%. He gives a talk on the system in April 2025, archived at PIRSA.

  7. February 2026

    The Aletheia math research agent

    In February 2026, he is among the authors of "Towards Autonomous Mathematics Research", which introduces Aletheia, a math research agent powered by an advanced version of Gemini Deep Think. The paper reports autonomous solutions to four open questions from Bloom's Erdős Conjectures database.

  8. As of October 2026

    Working at Google DeepMind

    As of October 2026, he works at Google DeepMind, with research interests that span machine learning, string theory and supergravity.

Career

  1. NowGoogle DeepMind

Education

  1. 2017University at Albany, State University of New YorkPhD

From the biography.

  1. 16% of AI career moves in the 2020s went from a university to a companyIncludes Yuri Chervonyi. We estimate that 16% of AI career moves in the 2020s went from a university to a company; 29% of moves starting at a university went to a company, and 76% ended the university role.
  2. Google had the most people shaping AI every year from 2014 to 2025Includes Google. Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.
  3. Google has the highest combined h-index among AI organizationsIncludes Google. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  4. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes Google. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.

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 Yuri Chervonyi yet. Send us a free-licensed photo.

  1. "Killing(-Yano) Tensors in String Theory" by Yuri Chervonyi and Oleg Lunin scholarsarchive.library.albany.edu. Checked 20 October 2026.
  2. "Hidden symmetries in string theory" by Iurii Chervonyi scholarsarchive.library.albany.edu. Checked 20 October 2026.
  3. [1710.08541] Hidden symmetries in string theory arxiv.org. Checked 20 October 2026.
  4. [2211.07357] Controlling Commercial Cooling Systems Using Reinforcement Learning arxiv.org. Checked 20 October 2026.
  5. [2211.07357] Controlling Commercial Cooling Systems Using Reinforcement Learning (ar5iv) ar5iv.labs.arxiv.org. Checked 20 October 2026.
  6. [2307.11546] Towards practical reinforcement learning for tokamak magnetic control arxiv.org. Checked 20 October 2026.
  7. AI achieves silver-medal standard solving International Mathematical Olympiad problems — Google DeepMind deepmind.google. Checked 20 October 2026.
  8. Gold-medalist Performance in Solving Olympiad Geometry with AlphaGeometry2 jmlr.org. Checked 20 October 2026.
  9. Yuri Chervonyi | PIRSA pirsa.org. Checked 20 October 2026.
  10. [2602.10177] Towards Autonomous Mathematics Research arxiv.org. Checked 20 October 2026.
  11. Yuri Chervonyi - Google Scholar scholar.google.com. Checked 20 October 2026.