PeopleofSuperintelligence

Researcher

Nika Haghtalab

Nika Haghtalab is a UC Berkeley computer scientist who builds the mathematical theory of machine learning and of its interactions with economics and society. She earned her Ph.D. at Carnegie Mellon in 2018, was a postdoctoral researcher at Microsoft Research New England in 2018–2019, and taught at Cornell in 2019–2020 before moving to Berkeley. Her 2022 paper with Michael Jordan and Eric Zhao, On-Demand Sampling: Learning Optimally from Multiple Distributions, won a NeurIPS Outstanding Paper award, and she received a Sloan Research Fellowship and an AI2050 Early Career Fellowship in 2024. As of October 2026, she is an Associate Professor with tenure in Berkeley's EECS department, co-directs the CLIMB center, and is a Program Chair of NeurIPS 2026.

Now at
Learning Theory Alliance and University of California, Berkeley
Badges

Timeline

  1. 2017

    Collaborative PAC learning

    With Avrim Blum, Ariel Procaccia and Mingda Qiao, she coauthors "Collaborative PAC Learning" for NIPS 2017. The paper introduces a model in which k players attempt to learn the same underlying concept.

  2. 2018–2019

    A Carnegie Mellon doctorate

    She earns a Ph.D. in computer science at Carnegie Mellon in 2018, co-advised by Avrim Blum and Ariel Procaccia. Her thesis, Foundation of Machine Learning, by the People, for the People, wins the CMU School of Computer Science Dissertation Award in 2018 and a SIGecom Dissertation Honorable Mention Award in 2019.

  3. 2018–2020

    Microsoft Research, then Cornell

    She is a postdoctoral researcher at Microsoft Research New England in 2018–2019. In 2019–2020, she is an assistant professor in the computer science department at Cornell University, before she comes to UC Berkeley.

  4. Since 2020

    Co-founding Learning Theory Alliance

    In 2020, she becomes a co-founder of the Learning Theory Alliance (LeT-All), a mentorship network with regular mentorship activities for the learning theory community.

  5. 2022

    A NeurIPS outstanding paper

    She, Michael Jordan and Eric Zhao win a NeurIPS 2022 Outstanding Paper award for On-Demand Sampling: Learning Optimally from Multiple Distributions. The paper establishes the optimal sample complexity of multi-distribution learning paradigms such as collaborative learning, group distributionally robust optimization and fair federated learning. In 2022, she also receives an NSF CAREER award.

  6. 2023–2024

    Jailbreaks and covert finetuning

    In July 2023, she coauthors Jailbroken: How Does LLM Safety Training Fail? with Alexander Wei and Jacob Steinhardt, which hypothesizes two failure modes of safety training and tests attacks against GPT-4 and Claude v1.3. In June 2024, she is one of six authors of Covert Malicious Finetuning, a method that makes a finetuned GPT-4 act on harmful instructions 99% of the time while avoiding detection.

  7. 2024

    Sloan and AI2050 fellowships

    In February 2024, she is one of nine Berkeley assistant professors named 2024 Sloan Research Fellows. In March 2024, she is among the 19 Schmidt Sciences AI2050 Early Career Fellows, who each receive up to $300,000 over two years; her project addresses social and strategic considerations in machine learning systems.

  8. As of October 2026

    Berkeley professor and NeurIPS chair

    She is an Associate Professor with tenure in the Department of Electrical Engineering and Computer Sciences at UC Berkeley, where she co-directs the CLIMB center and belongs to the Berkeley Artificial Intelligence Research (BAIR) Lab. She is a Program Chair of NeurIPS 2026 and, in Spring 2026, a visiting scientist and program organizer for the Simons Institute's Federated and Collaborative Learning program.

Career

  1. Since 2020Learning Theory AllianceCo-founder
  2. NowUniversity of California, BerkeleyAssociate Professor with tenure
  3. 2026Simons Institute for the Theory of ComputingVisiting Scientist and Program Organizer
  4. 2019–2020Cornell UniversityAssistant Professor
  5. 2018–2019Microsoft ResearchPostdoctoral Researcher

Education

  1. 2018Carnegie Mellon UniversityPh.D.

From the biography.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Nika Haghtalab. 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.
  2. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes University of California, Berkeley. 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.
  3. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes University of California, Berkeley. 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.
  4. The median h-index is 48 at universities and 14 at frontier AI labsIncludes University of California, Berkeley. 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.

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 Nika Haghtalab yet. Send us a free-licensed photo.

  1. Collaborative PAC Learning papers.nips.cc. Checked 22 October 2026.
  2. Nika Haghtalab | EECS at UC Berkeley www2.eecs.berkeley.edu. Checked 22 October 2026.
  3. Nika Haghtalab - Homepage people.eecs.berkeley.edu. Checked 22 October 2026.
  4. Nika Haghtalab | Research UC Berkeley vcresearch.berkeley.edu. Checked 22 October 2026.
  5. Nika Haghtalab | Simons Institute for the Theory of Computing simons.berkeley.edu. Checked 22 October 2026.
  6. Announcing the NeurIPS 2022 Awards blog.neurips.cc. Checked 22 October 2026.
  7. On-Demand Sampling: Learning Optimally from Multiple Distributions arxiv.org. Checked 22 October 2026.
  8. Jailbroken: How Does LLM Safety Training Fail? arxiv.org. Checked 22 October 2026.
  9. Covert Malicious Finetuning: Challenges in Safeguarding LLM Adaptation arxiv.org. Checked 22 October 2026.
  10. Nine young faculty members receive prized Sloan Research Fellowships news.berkeley.edu. Checked 22 October 2026.
  11. Statistics, computer science faculty among Schmidt Sciences AI2050 fellows cdss.berkeley.edu. Checked 22 October 2026.
  12. NeurIPS 2026 neurips.cc. Checked 22 October 2026.