PeopleofSuperintelligence

Researcher

Maarten Sap

Maarten Sap is a Carnegie Mellon assistant professor and AI2 AI safety lead whose research examines social intelligence and bias in language. He earned a degree in communications and information systems and a PhD in computer science and engineering, then coauthored ATOMIC, an atlas of commonsense reasoning. He coauthored the open Olmo 3 language models and Artificial Hivemind, which won NeurIPS's 2025 Best Paper (DB track) award. In 2025, he also received a Packard Fellowship.

Works
251
Citations
9,654
h-index
41

Publication numbers from OpenAlex, as of 15 October 2026.

Timeline

  1. 2010–2014

    A degree in information systems

    At École Polytechnique Fédérale de Lausanne, he earns a BS in Communications and Information Systems in 2014.

  2. 2015–2022

    A PhD in natural language processing

    At the University of Washington, he completes a PhD in computer science and engineering focused on natural language processing in 2022 after starting his studies in 2015.

  3. 2019

    ATOMIC commonsense reasoning

    In 2019, he coauthors ATOMIC, an atlas of everyday commonsense reasoning organized through 877,000 textual descriptions of inferential knowledge.

  4. 2021–2022

    An AI2 Young Investigator

    At AI2, he works as a postdoctoral researcher and Young Investigator from 2021 to 2022, then joins Carnegie Mellon as an assistant professor in 2022.

  5. 2025

    A Packard fellowship and paper prize

    He becomes a 2025 Packard Fellow at Carnegie Mellon. Artificial Hivemind, which he coauthors, wins NeurIPS's Best Paper (DB track) award in 2025.

  6. 2025

    Olmo 3 language models

    Team Olmo's 2025 preprint lists him among its authors and introduces fully open language models at 7B and 32B parameter scales.

  7. Since 2026

    AI2 safety lead

    Since 2026, he serves as AI2's Visiting Senior Research Scientist and AI Safety Lead while continuing as a Carnegie Mellon assistant professor. His research measures social and interactional intelligence in AI and studies inequality, safety risks, and socio-cultural biases in language.

Sources and credits

Photo: We have no published photo of Maarten Sap yet. Send us a free-licensed photo.

  1. Maarten Sap - CV maartensap.com. Checked 14 October 2026.
  2. [1811.00146] ATOMIC: An Atlas of Machine Commonsense for If-Then Reasoning arxiv.org. Checked 14 October 2026.
  3. Young investigators program | Ai2 allenai.org. Checked 14 October 2026.
  4. Maarten Sap • The David and Lucile Packard Foundation packard.org. Checked 14 October 2026.
  5. NeurIPS 2025 Awards neurips.cc. Checked 14 October 2026.
  6. [2512.13961] Olmo 3 arxiv.org. Checked 14 October 2026.
  7. Maarten Sap | Human-Computer Interaction Institute hcii.cmu.edu. Checked 14 October 2026.