PeopleofSuperintelligence

Researcher

Rylan Schaeffer

Rylan Schaeffer is a machine learning researcher who studies how to measure and predict what large language models can do. After studying computer science and statistics at UC Davis, he earned a research master's in cognitive neuroscience at University College London, worked as a data scientist at Uber, and completed a master's in computer science engineering at Harvard. His 2023 paper "Are Emergent Abilities of Large Language Models a Mirage?", written with Brando Miranda and Sanmi Koyejo, was named an outstanding main-track paper at NeurIPS 2023, and in 2024 he followed it with work on model collapse and on predicting downstream capabilities from scale. He worked on Gemini pretraining at Google from 2024 to 2025 and joined Meta in 2025. He is a Stanford computer science PhD student on leave, advised by Sanmi Koyejo, and says he is creating something new.

Now at
Meta
Works
61
Citations
1,153
h-index
13
Badges

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2011–2016

    Two degrees at UC Davis

    He earns two BS degrees at UC Davis, in computer science engineering and statistics, finishing in 2016. He begins in biomedical engineering, switches to business, and designs a major of his own on persuasion and motivation before settling on computer science and statistics.

  2. 2017–2018

    Metacognition research at UCL

    He earns a Master of Research in Cognitive Neuroscience at University College London's Institute of Cognitive Neuroscience, studying metacognition under Stephen Fleming. With Julius Kunze and Louis Kirsch, he starts UCL's Artificial Intelligence Journal Club, which hosts 25 speakers in 9 months.

  3. 2018–2019

    Data scientist at Uber

    He works as a data scientist on Uber's Intelligent Decision Systems team. The team specializes in time-series forecasting and anomaly detection.

  4. 2019–2021

    Harvard master's and MIT research

    He pursues a Master of Engineering in Computer Science Engineering at Harvard University and conducts research with Ila Fiete at MIT.

  5. 2023

    A NeurIPS outstanding paper

    In April 2023, he and Brando Miranda and Sanmi Koyejo post "Are Emergent Abilities of Large Language Models a Mirage?", which argues that apparent emergent abilities appear because of the researcher's choice of metric. NeurIPS lists it among its two Outstanding Main Track Papers in December 2023.

  6. 2024

    Predictable scaling and model collapse

    In April 2024, he coauthors "Is Model Collapse Inevitable?", which shows that accumulating synthetic data alongside the original real data avoids model collapse. In June 2024, he is first author of a paper showing that benchmark accuracy, computed from negative log likelihoods through a sequence of transformations, progressively loses its statistical relationship to scale, and the paper works toward scaling-predictable evaluations of frontier AI models.

  7. 2024–2025

    Gemini pretraining at Google

    He works as a research scientist on Google Gemini Pretraining & Data, a post that arrives with the Character AI acquihire. He contributes to the development of Gemini 2.5.

  8. As of October 2026

    Stanford PhD on leave after Meta

    He joins Meta in 2025 as an AI research scientist and works in its TBD lab. His profile now lists him as ex-Meta TBD and on leave from Stanford, where he is a computer science PhD student in Sanmi Koyejo's STAIR lab, and says he is creating something new.

Career

  1. From 2025MetaAI Research Scientist
  2. 2024–2025GoogleResearch Scientist
  3. 2021–2026Stanford University
  4. 2019–2021Massachusetts Institute of TechnologyResearcher
  5. 2018–2019UberData Scientist

Education

  1. 2021Harvard UniversityMaster of Engineering in Computer Science Engineering
  2. 2018University College LondonMaster of Research in Cognitive Neuroscience
  3. 2016University of California, DavisBS degrees
  4. Stanford UniversityPhD

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

Most-cited works

  1. Are Emergent Abilities of Large Language Models a Mirage?2023 · neural information processing systems · 196 citationsWith Sanmi Koyejo
  2. A brain-wide map of neural activity during complex behaviour2025 · Nature · 181 citationsWith Peter Dayan
  3. Brain-wide representations of prior information in mouse decision-making2025 · Nature · 120 citationsWith Peter Dayan
  4. DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models2023 · neural information processing systems · 104 citationsWith Bo Li and Sanmi Koyejo
  5. Lightning Pose: improved animal pose estimation via semi-supervised learning, Bayesian ensembling and cloud-native open-source tools2024 · Nature Methods · 87 citationsWith Peter Dayan

From OpenAlex, as of 21 October 2026.

  1. Of 943 people shaping AI who worked at Google, 446 work there nowIncludes Rylan Schaeffer. Our estimates of where the 943 people shaping AI who worked at Google and Google DeepMind work now: 446 at Google and 452 across 278 other organizations.
  2. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Rylan Schaeffer. 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.
  3. 16% of AI career moves in the 2020s went from a university to a companyIncludes Rylan Schaeffer. We estimate that 16% of AI career moves in the 2020s went from a university to a company; 29% of moves starting at a university went to a company, and 76% ended the university role.
  4. 23% of people shaping AI who left a university for a company later returned to a universityIncludes Rylan Schaeffer. We estimate that 468 of the 2,011 people shaping AI who left a university for a company (23%) later returned to a university after a median of 5 years, and 163 of 829 (20%) made the reverse round trip.

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 Rylan Schaeffer yet. Send us a free-licensed photo.

  1. About | Rylan Schaeffer rylanschaeffer.github.io. Checked 20 October 2026.
  2. [2304.15004] Are Emergent Abilities of Large Language Models a Mirage? arxiv.org. Checked 20 October 2026.
  3. Announcing the NeurIPS 2023 Paper Awards – NeurIPS Blog blog.neurips.cc. Checked 20 October 2026.
  4. [2404.01413] Is Model Collapse Inevitable? Breaking the Curse of Recursion by Accumulating Real and Synthetic Data arxiv.org. Checked 20 October 2026.
  5. [2406.04391] Why Has Predicting Downstream Capabilities of Frontier AI Models with Scale Remained Elusive? arxiv.org. Checked 20 October 2026.
  6. Home | Rylan Schaeffer rylanschaeffer.github.io. Checked 20 October 2026.
  7. Rylan Schaeffer | alphaXiv alphaxiv.org. Checked 20 October 2026.
  8. Rylan Schaeffer | Digg digg.com. Checked 20 October 2026.