PeopleofSuperintelligence

Researcher

Trieu Trinh

Also known as Trieu H. Trinh

Trieu Trinh is an AI researcher whose work spans language modeling, image representation, and mathematical reasoning. He coauthored a 2018 commonsense-reasoning paper and a 2019 self-supervised image-embedding method, then earned a PhD at NYU and helped develop AlphaGeometry, which solved 25 of 30 Olympiad geometry problems. In 2024 he led AlphaGeometry 2 for the IMO silver-level result; in 2025 he contributed to Google DeepMind’s Gemini Deep Think team, and in 2026 he coauthored a report on Aletheia’s FirstProof results. As of October 2026, he is a research scientist at Google DeepMind.

Badges

Timeline

  1. 2018

    Commonsense reasoning models

    In 2018, he coauthors “A Simple Method for Commonsense Reasoning” with Quoc V. Le. The paper uses unsupervised language models to score multiple-choice commonsense questions.

  2. 2019

    Selfie image embeddings

    In 2019, he coauthors Selfie, a self-supervised pretraining method for image embeddings. It masks image patches and trains a model to select the missing patch from distractors.

  3. 2024

    A PhD and AlphaGeometry

    At NYU, he completes a PhD in He He’s research group in 2024. In January 2024, he coauthors AlphaGeometry, which solves 25 of 30 Olympiad geometry problems within the standard time limit.

  4. July 2024

    Leading AlphaGeometry 2

    In July 2024, he leads AlphaGeometry 2 development with Yuri Chervonyi. Together with AlphaProof, the systems solve four of six IMO problems and score 28 points, matching silver-medal level.

  5. 2025

    Gemini Deep Think at IMO

    For the 2025 IMO, he contributes to Google DeepMind’s Gemini Deep Think team. Gemini Deep Think solves five of six problems for 35 points, a gold-medal-level result.

  6. November 2025

    IMO-Bench for mathematical reasoning

    In November 2025, he coauthors “Towards Robust Mathematical Reasoning,” which introduces IMO-Bench, an advanced benchmark suite for mathematical reasoning at IMO level.

  7. February 2026

    Aletheia at FirstProof

    In February 2026, he coauthors a report on Aletheia, a mathematics research agent powered by Gemini 3 Deep Think. Aletheia solves six of ten FirstProof problems by majority expert assessment.

  8. As of October 2026

    Google DeepMind research scientist

    As of October 2026, he is a research scientist at Google DeepMind.

Career

  1. NowGoogle DeepMindResearch scientist

Education

  1. 2024New York UniversityPhD

From the biography.

  1. 16% of AI career moves in the 2020s went from a university to a companyIncludes Trieu Trinh. 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 Trieu Trinh yet. Send us a free-licensed photo.

  1. A Simple Method for Commonsense Reasoning arxiv.org. Checked 21 October 2026.
  2. Selfie: Self-supervised Pretraining for Image Embedding arxiv.org. Checked 21 October 2026.
  3. Group - He He hhexiy.ai. Checked 21 October 2026.
  4. AlphaGeometry: An Olympiad-level AI system for geometry — Google DeepMind deepmind.google. Checked 21 October 2026.
  5. AI achieves silver-medal standard solving International Mathematical Olympiad problems — Google DeepMind deepmind.google. Checked 21 October 2026.
  6. Advanced version of Gemini with Deep Think officially achieves gold-medal standard at the International Mathematical Olympiad — Google DeepMind deepmind.google. Checked 21 October 2026.
  7. Towards Robust Mathematical Reasoning arxiv.org. Checked 21 October 2026.
  8. Aletheia tackles FirstProof autonomously arxiv.org. Checked 21 October 2026.