PeopleofSuperintelligence

Researcher

Catherine Olsson

Catherine Olsson is an AI researcher whose career spans machine-learning engineering, alignment and interpretability. She graduated from MIT in 2012, studied computational neuroscience at NYU, and moved from OpenAI to Google Brain. At Open Philanthropy she funded AI research, then became a research engineer at Anthropic and coauthored papers on helpful assistants and induction heads. As of October 2026, she works in program management at Anthropic.

Now at
Anthropic
Works
21
Citations
3,090
h-index
17
Badges

Publication numbers from OpenAlex, as of 19 October 2026.

Timeline

  1. 2012–2018

    From MIT to Google Brain

    She graduates from MIT in 2012 after studying computer and cognitive science, then studies at NYU's Center for Neural Science. In 2018, she works as a research software engineer at Google Brain after working as a software engineer at OpenAI.

  2. 2020

    Grantmaking for AI research

    At Open Philanthropy, she serves as a grantmaker and funds technical research that could help prepare for possible AI futures.

  3. 2021–2022

    Assistant alignment papers

    She coauthors A General Language Assistant as a Laboratory for Alignment (2021), Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback (2022), and Constitutional AI: Harmlessness from AI Feedback (2022). These papers test language-model training methods that use human preferences and AI feedback to shape assistant behavior.

  4. 2022

    A paper on induction heads

    She works as a research engineer at Anthropic and is the lead author of In-context Learning and Induction Heads, a 2022 mechanistic interpretability paper.

  5. 2023

    A study of human V1–V3

    She coauthors a 2023 study of orientation-tuned surround normalization in human V1–V3.

  6. As of October 2026

    Program management at Anthropic

    As of October 2026, she works in program management at Anthropic.

Career

  1. Since 2022AnthropicResearch Engineer
  2. From 2020Open PhilanthropyGrantmaker
  3. Until 2018OpenAISoftware Engineer
  4. From 2018Google BrainResearch Software Engineer

Education

  1. 2012Massachusetts Institute of Technology
  2. New York University

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

Most-cited works

  1. Dota 2 with Large Scale Deep Reinforcement Learning2019 · 1,063 citations
  2. Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback2022 · 417 citationsWith Chris Olah, Liane Lovitt and Zac Hatfield-Dodds
  3. AI-to-AI Code Reviews of GitHub Pull Requests2022 · 319 citationsWith Cameron McKinnon, Carol Chen, Chris Olah, Ethan Perez, Liane Lovitt and Samuel R. Bowman
  4. Predictability and Surprise in Large Generative Models2022 · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · 206 citationsWith Chris Olah, Liane Lovitt and Zac Hatfield-Dodds
  5. Discovering Language Model Behaviors with Model-Written Evaluations2023 · 188 citationsWith Cameron McKinnon, Chris Olah, Ethan Perez, Liane Lovitt, Samuel R. Bowman and Zac Hatfield-Dodds

From OpenAlex, as of 19 October 2026.

  1. 12 of 76 people shaping AI with a former OpenAI role now work at AnthropicIncludes Catherine Olsson. Our estimates of where OpenAI recruits came from and where people with former OpenAI roles work now.
  2. The median h-index is 54 at universities and 19 at frontier AI labsIncludes Catherine Olsson. The median h-index is 54 at universities and 19 at frontier AI labs. We compare publishing careers, citations and moves to estimate how universities and frontier labs contribute to AI research.
  3. Of 435 people shaping AI who worked at Google, 178 work there nowIncludes Catherine Olsson. Our estimates of where former Google and Google DeepMind people work now, including moves within Google.
  4. Nonprofit roles are 2.3 times as common in AI safetyIncludes Anthropic. We estimate where people focused on safety, alignment and evaluation work. Nonprofit roles appear for 26.8% of safety people with known affiliations, compared with 11.5% in other areas.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

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

  1. Catherine Olsson, MIT computer + cognitive science esp.mit.edu. Checked 18 October 2026.
  2. People – Winawer Lab wp.nyu.edu. Checked 18 October 2026.
  3. PhD or programming? Fast paths into aligning AI as a machine learning engineer, according to ML engineers Catherine Olsson & Daniel Ziegler 80000hours.org. Checked 18 October 2026.
  4. Catherine Olsson ’08 - Lakeside School lakesideschool.org. Checked 18 October 2026.
  5. A General Language Assistant as a Laboratory for Alignment arxiv.org. Checked 18 October 2026.
  6. Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback arxiv.org. Checked 18 October 2026.
  7. Constitutional AI: Harmlessness from AI Feedback arxiv.org. Checked 18 October 2026.
  8. BlackboxNLP 2022 | Analyzing and interpreting neural networks for NLP blackboxnlp.github.io. Checked 18 October 2026.
  9. Normalization by orientation-tuned surround in human V1-V3 journals.plos.org. Checked 18 October 2026.
  10. Claude Founder House San Francisco | Anthropic anthropic.com. Checked 18 October 2026.