PeopleofSuperintelligence

Researcher

Stefano Ermon

Also known as S. Ermon

Stefano Ermon is a Stanford computer scientist whose work spans distribution comparison, generative diffusion, and language-model alignment. He earned electrical-engineering degrees at Padova and a computer science Ph.D. at Cornell, then helped launch Atlas AI for development data. Papers he coauthored received ICLR Outstanding Paper Awards in 2021 and 2022, a NeurIPS 2023 runner-up award, and ICML’s 2024 Best Paper Award. As of October 2026, he is a Stanford associate professor; in September 2026, he announced Inception’s Mercury 2.5 diffusion language model as CEO.

Now at
Stanford University and Inception
Works
485

Publication numbers from OpenAlex, as of 14 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.

Photo of Stefano Ermon

Timeline

  1. 2006–2015

    From Padova to Cornell

    At Padova, he earns B.Sc. and M.Sc. degrees in electrical engineering in 2006 and 2008. During his Cornell computer science Ph.D., he accepts a tenure-track assistant professor position at Stanford; he completes the degree in 2015.

  2. 2018

    An AI research award

    He receives the IJCAI Computers and Thought Award.

  3. March 2019

    Launching Atlas AI

    He helps launch Atlas AI with Stanford professors David Lobell and Marshall Burke to give decision-makers in developing countries access to low-cost data.

  4. 2021

    A score-based framework

    With Yang Song and coauthors, he publishes Score-Based Generative Modeling through Stochastic Differential Equations. The paper uses forward and reverse stochastic differential equations for generative modeling and wins an ICLR Outstanding Paper Award.

  5. 2022

    Comparing distributions for decisions

    With coauthors, he proposes a way to compare probability distributions by the optimal loss for a decision task. The paper wins an ICLR Outstanding Paper Award.

  6. 2023

    Aligning models with preferences

    He coauthors Direct Preference Optimization (DPO), a single-stage method for aligning language models with human preferences. NeurIPS names it an Outstanding Main Track Runner-Up.

  7. 2024

    Discrete diffusion for text

    With Aaron Lou and Chenlin Meng, he develops SEDD, which uses score entropy for discrete diffusion in language modeling. The paper wins ICML’s 2024 Best Paper Award.

  8. As of October 2026

    A language model in production

    At Stanford, he is an associate professor of computer science. In September 2026, he announces Mercury 2.5 as Inception’s CEO; the model is a diffusion language model.

Sources and credits

Photo: cs.stanford.edu; Stefano Ermon; creator not stated. License: Not stated by source. Source: cs.stanford.edu. Cropped to a square and resized. Ask us to remove or replace this photo.

  1. Stefano Ermon cs.cornell.edu. Checked 14 October 2026.
  2. Speaker Profiles | College of Computing and Data Science | NTU Singapore ntu.edu.sg. Checked 14 October 2026.
  3. IJCAI Awards ijcai.org. Checked 14 October 2026.
  4. The Rockefeller Foundation Establishes Atlas AI - New Startup to Generate Actionable Intelligence on Global Development Challenges prnewswire.com. Checked 14 October 2026.
  5. Generative Modeling by Estimating Gradients of the Data Distribution yang-song.net. Checked 14 October 2026.
  6. Score-Based Generative Modeling through Stochastic Differential Equations arxiv.org. Checked 14 October 2026.
  7. Announcing the ICLR 2022 Outstanding Paper Award Recipients blog.iclr.cc. Checked 14 October 2026.
  8. Announcing the NeurIPS 2023 Paper Awards blog.neurips.cc. Checked 14 October 2026.
  9. Congratulations to Aaron Lou, Chenlin Meng, and Stefano Ermon for an ICML 2024 Best Paper Award! ai.stanford.edu. Checked 14 October 2026.
  10. Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution proceedings.mlr.press. Checked 14 October 2026.
  11. Stefano Ermon cs.stanford.edu. Checked 14 October 2026.
  12. Introducing Mercury 2.5 inceptionlabs.ai. Checked 14 October 2026.