PeopleofSuperintelligence

Researcher

Yinzhan Xu

Yinzhan Xu is a theoretical computer scientist whose work connects AI with mathematical research. At MIT, he worked on Coq-based linear algebra as an undergraduate, competed on its ICPC team in 2019, and earned a B.S. in computer science and math and a PhD under Virginia Vassilevska Williams. His 2024 paper with Ce Jin won the Danny Lewin Best Student Paper Award at STOC, and his 2025 work applied private expander decomposition to graph sparsification. In March 2026, he coauthored lower bounds for transformer computation; in June 2026, ChatGPT 5.5 Pro discovered the initial proof in his Furthest Pair paper, which he and his coauthors validated and edited. He is a postdoctoral researcher in computer science at UCSD, hosted by Barna Saha.

Timeline

  1. 2017–2018

    Coq research at MIT

    In 2017–2018, he is an MIT EECS Undergraduate Research and Innovation Scholar under Adam Chlipala, working on deductive synthesis of linear algebra formulas in the Coq proof assistant.

  2. April 2019

    ICPC World Finals second place

    He competes on MIT's 2019 ICPC team in Porto, which places second globally at the World Finals.

  3. 2020–2023

    Fast algorithms and lower bounds

    From 2020 to 2023, he coauthors work on min-plus products, shortest paths and lower bounds for dynamic problems, with papers appearing at SODA, FOCS, STOC and ICALP.

  4. 2024

    A STOC paper wins an award

    At STOC in 2024, he and Ce Jin coauthor Shaving Logs via Large Sieve Inequality: Faster Algorithms for Sparse Convolution and More, which earns the Danny Lewin Best Student Paper Award.

  5. 2025

    Private graph sparsification

    At ICML in 2025, he coauthors Breaking the n^1.5 Additive Error Barrier for Private and Efficient Graph Sparsification via Private Expander Decomposition.

  6. By 2026

    MIT bachelor's and PhD

    By 2026, he completes a PhD at MIT under Virginia Vassilevska Williams and earns a B.S. in computer science and math there.

  7. 2026

    Transformer bounds and AI proof

    In March 2026, he and Barna Saha, Christopher Ye and Hantao Yu submit On the Computational Hardness of Transformers, which establishes lower bounds for multi-head, multi-layer transformers. In June 2026, he, Barna Saha and Christopher Ye submit Furthest Pair Requires Quadratic Time in Superconstant Dimension under SETH; ChatGPT 5.5 Pro initially discovers the proof, which the authors validate and substantially edit.

  8. As of October 2026

    Postdoctoral research at UCSD

    As of October 2026, he is a postdoc in UCSD's Computer Science department, hosted by Barna Saha. His research focuses on theoretical computer science, especially fine-grained complexity and algorithms.

Career

  1. NowUniversity of California, San DiegoPostdoctoral researcher
  2. 2017–2018Massachusetts Institute of TechnologyUndergraduate Research and Innovation Scholar

Education

  1. Massachusetts Institute of TechnologyB.S.

From the biography.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes University of California, San Diego. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  2. Universities keep AI researchers for a median of 6 years, compared with 2 years at frontier labsIncludes University of California, San Diego. Across 15,811 dated work stays (6,529 people), median role tenure runs from 6 years at universities to 2 years at frontier labs. We compare all five organization types, starting decades from the 1970s to the 2020s, and the organizations that keep people longest.
  3. The median h-index is 48 at universities and 14 at frontier AI labsIncludes University of California, San Diego. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.
  4. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Massachusetts Institute of Technology. We 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 Yinzhan Xu yet. Send us a free-licensed photo.

  1. Yinzhan Xu - MIT SuperUROP – Advanced Undergraduate Research Opportunities Program superurop.mit.edu. Checked 20 October 2026.
  2. MIT team places second at ACM’s global programming competition | MIT CSAIL csail.mit.edu. Checked 20 October 2026.
  3. Yinzhan Xu xuyinzhan.github.io. Checked 20 October 2026.
  4. [2603.11332] On the Computational Hardness of Transformers arxiv.org. Checked 20 October 2026.
  5. [2606.25887] Furthest Pair Requires Quadratic Time in Superconstant Dimension under SETH arxiv.org. Checked 20 October 2026.