PeopleofSuperintelligence

Researcher

Rowan Zellers

Rowan Zellers is an AI researcher known for commonsense benchmarks and multimodal models that learn from video. During his University of Washington PhD under Yejin Choi and Ali Farhadi, he coauthored the SWAG and HellaSwag datasets, was lead author of the Grover fake-news paper, and shared a NeurIPS 2021 Outstanding Paper award for MAUVE. His MERLOT and MERLOT Reserve models learned from millions of YouTube videos. After finishing his PhD in 2022 he chose OpenAI, where he worked on GPT-4 and led video perception for GPT-4o in 2024. He is a researcher at Thinking Machines, working on vision and large language models.

Badges

Timeline

  1. 2016

    Harvey Mudd and a UW PhD

    He graduates from Harvey Mudd College in 2016 with majors in computer science and mathematics. His PhD at the University of Washington, advised by Yejin Choi and Ali Farhadi, runs from 2016 to 2022.

  2. 2018–2019

    SWAG and HellaSwag datasets

    In 2018, he coauthors the SWAG dataset of 113k multiple choice questions about grounded situations with Yonatan Bisk, Roy Schwartz and Yejin Choi. In 2019, he coauthors HellaSwag with Ari Holtzman, Yonatan Bisk, Ali Farhadi and Yejin Choi, a commonsense challenge dataset on which state-of-the-art models struggle.

  3. 2019

    Grover and neural fake news

    He is lead author of the NeurIPS 2019 paper "Defending Against Neural Fake News", which presents Grover, a model that writes an article from a headline. Grover raises detection of AI-written news from 73% to 92% accuracy, and he argues for releasing the model to researchers so that defenses can improve.

  4. 2021

    A NeurIPS Outstanding Paper

    MAUVE, a comparison measure for open-ended text generation, receives a NeurIPS 2021 Outstanding Paper award. He coauthors it with Krishna Pillutla, Swabha Swayamdipta, John Thickstun, Sean Welleck, Yejin Choi and Zaid Harchaoui, and it compares the distribution of a text generation model with the distribution of human-written text.

  5. 2021–2022

    MERLOT learns from YouTube

    In 2021, he is first author of MERLOT, a model that learns multimodal script knowledge from millions of YouTube videos with transcribed speech. In 2022, he is first author of MERLOT Reserve, which learns from audio, subtitles and video frames and pretrains on 20 million YouTube videos.

  6. 2022

    Joining OpenAI after the PhD

    He finishes his PhD at the University of Washington in June 2022. After a job search that brings academic offers, he signs with OpenAI.

  7. 2023–2024

    GPT-4 and GPT-4o at OpenAI

    At OpenAI, he works on text and image understanding and evaluation for GPT-4 in 2023. For GPT-4o in 2024, he leads video perception, including architecture, pretraining, data and evaluations, and helps make a real-time audio and video demo happen.

  8. As of October 2026

    Researcher at Thinking Machines

    He is a researcher at Thinking Machines working on vision and large language models across the stack, and lives in San Francisco.

Career

  1. NowThinking Machines LabResearcher
  2. Until 2024OpenAI

Education

  1. 2022University of WashingtonPhD
  2. 2016Harvey Mudd CollegeUndergraduate student

From the biography.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Rowan Zellers. 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. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes Rowan Zellers. 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.
  3. 22 of 133 people shaping AI with a former OpenAI role now work at AnthropicIncludes Rowan Zellers. Our estimates of where OpenAI recruits came from and where 133 people with former OpenAI roles work now, including the companies and labs they founded.
  4. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Rowan Zellers. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.

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 Rowan Zellers yet. Send us a free-licensed photo.

  1. Rowan Zellers' Homepage rowanzellers.com. Checked 22 October 2026.
  2. Rowan Zellers (U. of Washington) - Grounding Language by Seeing, Hearing, and Interacting | DSI datascience.uchicago.edu. Checked 22 October 2026.
  3. Why I chose OpenAI over academia: reflections on the CS academic and industry job markets (part 2) - Rowan Zellers' blog rowanzellers.com. Checked 22 October 2026.
  4. SWAG: A Large-Scale Adversarial Dataset for Grounded Commonsense Inference - ACL Anthology aclanthology.org. Checked 22 October 2026.
  5. HellaSwag: Can a Machine Really Finish Your Sentence? - ACL Anthology aclanthology.org. Checked 22 October 2026.
  6. Allen School and AI2 researchers unveil Grover, a new tool for fighting fake news in the age of AI news.cs.washington.edu. Checked 22 October 2026.
  7. To detect fake news, this AI first learned to write it | TechCrunch techcrunch.com. Checked 22 October 2026.
  8. NeurIPS 2021 Awards neurips.cc. Checked 22 October 2026.
  9. MERLOT: Multimodal Neural Script Knowledge Models arxiv.org. Checked 22 October 2026.
  10. MERLOT Reserve: Neural Script Knowledge through Vision and Language and Sound arxiv.org. Checked 22 October 2026.
  11. Reflections on the CS academic and industry job markets (part 1) - Rowan Zellers' blog rowanzellers.com. Checked 22 October 2026.