PeopleofSuperintelligence

Researcher

Kawin Ethayarajh

Kawin Ethayarajh is a computer scientist whose behavioral machine learning research connects AI with economics and behavioral science. After founding the Review of Undergraduate Computer Science at the University of Toronto in 2015, he completed a computer science PhD at Stanford under Dan Jurafsky, where his work on dataset difficulty won an ICML Outstanding Paper award in 2022 and led to the Stanford Human Preferences dataset. In 2024, he introduced Kahneman-Tversky Optimization (KTO), a widely used algorithm for aligning language models with feedback, and served as a postdoctoral fellow at Princeton Language and Intelligence before joining the University of Chicago Booth School of Business in July 2025. He is an Assistant Professor of Applied AI and Kathryn and Grant Swick Faculty Scholar at Booth, where his 2026 study with Giulio Frey on mecha-nudges measures how Etsy listings changed after ChatGPT's release.

Timeline

  1. 2015–2017

    Founding RUCS at Toronto

    In 2015, he founds the Review of Undergraduate Computer Science (RUCS), an undergraduate journal at the University of Toronto. A computer science major at Victoria College, he serves on Governing Council's University Affairs Board and is named the 2017 recipient of the John H. Moss Scholarship for graduating students.

  2. 2018

    Random walk sentence embeddings

    In 2018, he publishes the sole-author paper 'Unsupervised Random Walk Sentence Embeddings: A Strong but Simple Baseline' in the proceedings of the Third Workshop on Representation Learning for NLP.

  3. 2019

    Geometry of contextual embeddings

    In 2019, his single-author paper 'How Contextual are Contextualized Word Representations?' appears at EMNLP-IJCNLP in Hong Kong. It compares the geometry of BERT, ELMo and GPT-2 embeddings and finds that contextual representations are anisotropic, occupying a narrow cone in vector space.

  4. 2021

    A Meta Fellowship

    He receives a Meta Fellowship in 2021 and also holds an NSERC PGS-D.

  5. 2022

    An ICML Outstanding Paper

    His paper 'Understanding Dataset Difficulty with V-Usable Information,' written with Yejin Choi and Swabha Swayamdipta, receives an Outstanding Paper award at ICML in 2022. It frames dataset difficulty as the lack of V-usable information and leads to Stanford Human Preferences (SHP), a dataset of 385K human preferences over Reddit responses. SHP is used to post-train Llama-2.

  6. 2024

    Kahneman-Tversky Optimization

    In 2024, he introduces Kahneman-Tversky Optimization (KTO) in 'KTO: Model Alignment as Prospect Theoretic Optimization,' written with Winnie Xu, Niklas Muennighoff, Dan Jurafsky and Douwe Kiela and presented at ICML 2024. KTO learns only from a binary signal of whether an output is desirable, matches or exceeds preference-based methods at scales from 1B to 30B, and is a widely used algorithm for aligning language models with feedback.

  7. 2024–2025

    Postdoc at Princeton Language and Intelligence

    After earning a B.Sc. and M.Sc. in computer science as a National Scholar at the University of Toronto and a PhD in computer science at Stanford advised by Dan Jurafsky, he serves as a postdoctoral fellow at Princeton Language and Intelligence in 2024 and 2025. At Princeton, he works on behavioral machine learning and making post-training with real-world feedback more efficient, drawing on behavioral economics research on how people perceive probability.

  8. Since 2025

    Assistant professor at Chicago Booth

    He joins the University of Chicago Booth School of Business in July 2025 from Princeton as Assistant Professor of Applied AI and Kathryn and Grant Swick Faculty Scholar, working on behavioral machine learning, a field at the intersection of AI, economics and behavioral science. In September 2025, he coauthors 'Humanline: Online Alignment as Perceptual Loss' with Sijia Liu and Niklas Muennighoff, proposing alignment objectives that incorporate perceptual distortions of probability into DPO, KTO and GRPO. In 2026, he and Giulio Frey introduce mecha-nudges, changes to how choices are presented that systematically influence AI agents, and find across more than six million Etsy listings that machine-usable information increases by 0.143 bits after ChatGPT's release.

Career

  1. Since 2025University of ChicagoAssistant Professor of Applied AI
  2. 2025Princeton UniversityPostdoctoral researcher

Education

  1. University of TorontoB.Sc.
  2. Stanford UniversityPhD

From the biography.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes University of Chicago. 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. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes University of Chicago. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  3. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes University of Chicago. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.
  4. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes University of Chicago. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.

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 Kawin Ethayarajh yet. Send us a free-licensed photo.

  1. RUCS | Review of Undergraduate Computer Science rucs-uoft.github.io. Checked 22 October 2026.
  2. Kawin Ethayarajh wins the John H. Moss Scholarship for graduating students | University of Toronto Alumni alumni.utoronto.ca. Checked 22 October 2026.
  3. Unsupervised Random Walk Sentence Embeddings: A Strong but Simple Baseline - ACL Anthology aclanthology.org. Checked 22 October 2026.
  4. Kawin Ethayarajh | Assistant Professor of Applied AI at UChicago Booth, working on behavior-bound machine learning kawine.github.io. Checked 22 October 2026.
  5. How Contextual are Contextualized Word Representations? Comparing the Geometry of BERT, ELMo, and GPT-2 Embeddings - ACL Anthology aclanthology.org. Checked 22 October 2026.
  6. Kawin Ethayarajh | The University of Chicago Booth School of Business chicagobooth.edu. Checked 22 October 2026.
  7. ICML 2022 Awards icml.cc. Checked 22 October 2026.
  8. stanfordnlp/SHP · Datasets at Hugging Face huggingface.co. Checked 22 October 2026.
  9. KTO: Model Alignment as Prospect Theoretic Optimization arxiv.org. Checked 22 October 2026.
  10. Events | Inside Our Program | Master of Science in Information Technology | Northwestern Engineering mccormick.northwestern.edu. Checked 22 October 2026.
  11. AI researcher Kawin Ethayarajh is redefining how AI learns from human behavior - CO/AI getcoai.com. Checked 22 October 2026.
  12. New Faculty Advance AI - Center for Applied Artificial Intelligence | Chicago Booth chicagobooth.edu. Checked 22 October 2026.
  13. [2509.24207] Humanline: Online Alignment as Perceptual Loss arxiv.org. Checked 22 October 2026.
  14. The Quiet Rewrite - Center for Applied Artificial Intelligence | Chicago Booth chicagobooth.edu. Checked 22 October 2026.
  15. Mecha-nudges for Machines arxiv.org. Checked 22 October 2026.