PeopleofSuperintelligence

Researcher

Xiaomeng Yang

Xiaomeng Yang is a Google DeepMind researcher who works on reinforcement learning and AI systems for mathematical reasoning. At Meta AI in 2022, they coauthored the Nocturne multi-agent driving benchmark and the AutoCAT framework for exploring cache-timing attacks, followed by the MACTA paper at ICLR 2023 and the Megalodon language model architecture in April 2024. They contributed to AlphaGeometry 2, which helped reach silver-medalist level at the July 2024 International Mathematical Olympiad and achieved an 84% solving rate on Olympiad geometry problems in a 2025 paper, and to the Gemini Deep Think teams that reached gold-medal level at the July 2025 IMO and the September 2025 ICPC World Finals. As of October 2026, they research artificial intelligence, reinforcement learning and autonomous mathematics at Google DeepMind, where they coauthored the 2026 paper introducing the Aletheia research agent.

Works
1
Citations
2
h-index
1

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2022

    Meta AI, Nocturne and AutoCAT

    While at Meta AI in 2022, Yang coauthors Nocturne, a scalable 2D driving benchmark for multi-agent learning under partial observability published in the NeurIPS 2022 Datasets and Benchmarks Track. In 2022, with researchers from Cornell University, Virginia Tech, Intel Corporation and Meta AI, they also coauthor AutoCAT, which uses reinforcement learning to explore cache-timing attacks and discovers the StealthyStreamline attack.

  2. 2023

    MACTA at ICLR 2023

    With Jiaxun Cui, Mulong Luo, Geunbae Lee and Peter Stone in 2023, they coauthor MACTA, an ICLR 2023 paper that applies multi-agent reinforcement learning to cache-timing attacks and detection.

  3. April 2024

    The Megalodon model architecture

    In April 2024, Yang and Xuezhe Ma coauthor Megalodon, a neural architecture designed for efficient large language model pretraining and inference with unlimited context length.

  4. July 2024

    AlphaGeometry 2 at the IMO

    AlphaGeometry 2, led by Trieu Trinh and Yuri Chervonyi, counts Yang among its key contributors in July 2024. Combined with AlphaProof, the system solves four of six International Mathematical Olympiad problems in July 2024, reaching the level of a silver medalist.

  5. 2025

    AlphaGeometry2 results in JMLR

    The 2025 Journal of Machine Learning Research paper on AlphaGeometry2 lists Yang as a coauthor alongside Yuri Chervonyi, Trieu H. Trinh, Quoc V. Le and Thang Luong. It reports an 84% solving rate on Olympiad geometry problems from the last 25 years, up from 54% previously.

  6. July 2025

    Gold at the 2025 IMO

    Gemini Deep Think solves five of six problems perfectly for 35 total points and reaches gold-medal level at the International Mathematical Olympiad in July 2025. Google DeepMind names Yang among the leads and key contributors to its IMO 2025 team.

  7. September 2025

    Gemini at the ICPC finals

    At the September 2025 ICPC World Finals, Gemini 2.5 Deep Think solves 10 of 12 problems and reaches gold-medal level. Google DeepMind names Yang among the leads and key contributors to the ICPC 2025 team.

  8. As of October 2026

    Aletheia and Google DeepMind research

    As of October 2026, they work on artificial intelligence, reinforcement learning and mathematical AI research at Google DeepMind. They coauthor a 2026 paper introducing Aletheia, a mathematics research agent built on Gemini Deep Think that generates, verifies and revises solutions in natural language.

  1. 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.
  2. 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.
  3. 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.
  4. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Google. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.

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 Xiaomeng Yang yet. Send us a free-licensed photo.

  1. Nocturne: a scalable driving benchmark for bringing multi-agent learning one step closer to the real world proceedings.neurips.cc. Checked 22 October 2026.
  2. AutoCAT: Reinforcement Learning for Automated Exploration of Cache-Timing Attacks ar5iv.labs.arxiv.org. Checked 22 October 2026.
  3. ICLR Poster MACTA: A Multi-agent Reinforcement Learning Approach for Cache Timing Attacks and Detection iclr.cc. Checked 22 October 2026.
  4. Megalodon: Efficient LLM Pretraining and Inference with Unlimited Context Length arxiv.org. Checked 22 October 2026.
  5. AI achieves silver-medal standard solving International Mathematical Olympiad problems deepmind.google. Checked 22 October 2026.
  6. Gold-medalist Performance in Solving Olympiad Geometry with AlphaGeometry2 jmlr.org. Checked 22 October 2026.
  7. Advanced version of Gemini with Deep Think officially achieves gold-medal standard at the International Mathematical Olympiad deepmind.google. Checked 22 October 2026.
  8. Gemini achieves gold-medal level at the International Collegiate Programming Contest World Finals deepmind.google. Checked 22 October 2026.
  9. Xiaomeng Yang - Google Scholar scholar.google.com. Checked 22 October 2026.
  10. Towards Autonomous Mathematics Research arxiv.org. Checked 22 October 2026.