PeopleofSuperintelligence

Researcher

David Persson

David Persson is a numerical analyst whose matrix algorithms help make deep-learning training more efficient. At EPFL, he coauthored trace-estimation variants of Hutch++ in 2021 and 2022 and completed a mathematics PhD under Daniel Kressner in June 2024. In September 2024, he joined the Flatiron Institute's Center for Computational Mathematics as a Flatiron Research Fellow and became a Courant Instructor at New York University. With collaborators, he developed The Polar Express for the Muon optimizer in May 2025, won a second prize in the IMA Leslie Fox Prize for Numerical Analysis in June 2025, and received an ICLR 2026 Outstanding Paper Honorable Mention in April 2026. As of October 2026, he is a Courant Instructor at NYU's Courant Institute of Mathematical Sciences and a Flatiron Research Fellow in the Center for Computational Mathematics.

Timeline

  1. 2021–2022

    Hutch++ trace estimation variants

    In September 2021, he coauthors "Improved variants of the Hutch++ algorithm for trace estimation" with Alice Cortinovis and Daniel Kressner, revising the work in May 2022.

  2. June 2024

    A PhD at EPFL

    He completes a mathematics PhD at EPFL in June 2024 under advisor Daniel Kressner. His doctoral thesis is titled "Randomized low-rank approximation and its applications."

  3. September 2024

    Flatiron and Courant roles

    In September 2024, he joins the Flatiron Institute's Center for Computational Mathematics as a Flatiron Research Fellow. He also serves as a Courant Instructor at New York University's Courant Institute of Mathematical Sciences.

  4. February 2025

    Block-Krylov methods for matrix functions

    With Tyler Chen and Christopher Musco, he posts "Randomized block-Krylov subspace methods for low-rank approximation of matrix functions" in February 2025.

  5. May 2025

    Polar Express for Muon

    With Noah Amsel, Christopher Musco, and Robert Gower, he introduces Polar Express in May 2025, a GPU-friendly method for computing the polar decomposition with matrix-matrix multiplications. Integrated into the Muon optimizer, the method yields consistent validation-loss improvements for a GPT-2 model trained on the FineWeb dataset.

  6. June 2025

    A Leslie Fox second prize

    He wins a second prize in the IMA Leslie Fox Prize for Numerical Analysis in June 2025 at the University of Strathclyde for his work on low-rank approximations of matrix functions.

  7. April 2026

    An ICLR honorable mention

    In April 2026, ICLR gives The Polar Express an Outstanding Paper Honorable Mention. Persson coauthors the paper with Noah Amsel, Christopher Musco, and Robert M. Gower, using approximation theory to design optimal polynomial approximations for the polar decomposition used by Muon.

  8. As of October 2026

    Courant and Flatiron research

    He is a Courant Instructor at NYU's Courant Institute of Mathematical Sciences and a Flatiron Research Fellow in the Center for Computational Mathematics. In August 2026, he works with Robert Gower and Noah Amsel to find optimal polynomial coefficients that update at every iteration and become a core component of the Muon optimizer.

Career

  1. Since 2024Simons FoundationFlatiron Research Fellow
  2. NowCourant Institute of Mathematical SciencesCourant Instructor

Education

  1. 2024EPFLPhD

From the biography.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes New York University. 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.
  2. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes New York University. 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.
  3. Direct Form 990 pay in the latest returns of 39 nonprofits with AI people on this site is highest at $2.1M at the Allen Institute for AI (Ai2) (2024)Includes Simons Foundation. IRS Form 990 compensation across 39 nonprofits with AI people on this site from 2012 to 2025: in the latest returns, Mozilla reports $6.1M including related-company pay (2024) and the Allen Institute for AI (Ai2) leads direct nonprofit pay at $2.1M (2024).
  4. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes New York University. Our estimates of the organizations where 550 people worked directly before joining 9 frontier AI labs across 620 career transitions, comparing the main feeder organizations per lab.

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 David Persson yet. Send us a free-licensed photo.

  1. Improved variants of the Hutch++ algorithm for trace estimation arxiv.org. Checked 22 October 2026.
  2. David Persson | Courant Instructor at NYU Courant & Flatiron Research Fellow davpersson.github.io. Checked 22 October 2026.
  3. Randomized low-rank approximation and its applications - EPFL memento.epfl.ch. Checked 22 October 2026.
  4. David Persson simonsfoundation.org. Checked 22 October 2026.
  5. Randomized block-Krylov subspace methods for low-rank approximation of matrix functions arxiv.org. Checked 22 October 2026.
  6. The Polar Express: Optimal Matrix Sign Methods and Their Application to the Muon Algorithm arxiv.org. Checked 22 October 2026.
  7. IMA Leslie Fox Prize for Numerical Analysis ima.org.uk. Checked 22 October 2026.
  8. Announcing the ICLR 2026 Outstanding Papers blog.iclr.cc. Checked 22 October 2026.
  9. Noah Amsel, Professor David Persson, and Professor Christopher Musco Receive ICLR 2026 Outstanding Paper Honorable Mention cims.nyu.edu. Checked 22 October 2026.
  10. How Flatiron Institute Mathematicians Helped Make AI More Efficient simonsfoundation.org. Checked 22 October 2026.