PeopleofSuperintelligence

Researcher

Valentina Pyatkin

Valentina Pyatkin is a computer scientist whose work connects language understanding with open language models. Her research spans semantic roles, abstract instructions, model post-training and precise instruction following. At AI2 and the University of Washington, she contributed to OLMo and coauthored Tülu 3; OLMo won ACL's 2024 Best Theme Paper Award, and her coauthored Political Compass or Spinning Arrow? appeared among ACL's Outstanding Papers. In 2025, she coauthored IFBench and Olmo 3. As of October 2026, she is a researcher at ETH AI Center, where she works on post-training for the Swiss AI Initiative and contributes to Apertus.

Timeline

  1. 2021

    Semantic role question generation

    Her academic path includes a BA from the University of Zurich and an MSc in artificial intelligence with distinction from the University of Edinburgh, followed by a computer science PhD at Bar Ilan University's NLP lab under Ido Dagan and Reut Tsarfaty. In 2021, she coauthors Asking It All, a two-stage model that generates contextual questions for semantic roles.

  2. 2022

    Grounding abstract instructions

    In 2022, she coauthors Draw Me a Flower, which introduces Hexagons, a 2D instruction-following game. The team derives an instruction-to-execution task from more than 4,000 visually grounded instructions about abstraction.

  3. 2022

    AI2 postdoctoral research

    In 2022, she is a postdoctoral researcher and Young Investigator at AI2 and the University of Washington, advised by Yejin Choi, and receives Schmidt Sciences' Israeli Women's Postdoctoral Award.

  4. 2024

    Open language modeling with OLMo

    In 2024, she contributes to OLMo at AI2 and the University of Washington, coauthoring a truly open language model released with open training data and training and evaluation code. The paper wins ACL's Best Theme Paper Award. Her coauthored Political Compass or Spinning Arrow? appears among ACL's Outstanding Papers.

  5. November 2024

    Tülu 3 post-training

    In November 2024, she coauthors Tülu 3 at AI2, a family of fully open post-trained models released with its data, code and training recipes.

  6. 2025

    IFBench for instruction following

    In 2025, she coauthors Generalizing Verifiable Instruction Following, accepted to NeurIPS 2025; it introduces IFBench to test precise instruction-following generalization on 58 new verifiable constraints. The study reports that reinforcement learning with verifiable rewards improves instruction following.

  7. December 2025

    Olmo 3's open model family

    In December 2025, she coauthors Olmo 3, a fully open family of 7B and 32B language models built for long-context reasoning, coding and instruction following.

  8. As of October 2026

    Research at ETH AI Center

    As of October 2026, she is a researcher at ETH AI Center, where she works on post-training of LLMs for the Swiss AI Initiative and contributes to Apertus.

Career

  1. NowETH ZurichResearcher
  2. From 2022Allen Institute for AI (Ai2)Postdoctoral researcher and Young Investigator
  3. From 2022University of WashingtonPostdoctoral researcher and Young Investigator

Education

  1. University of ZurichBA
  2. University of EdinburghMSc
  3. Bar-Ilan UniversityPhD

From the biography.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes ETH Zurich. 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 ETH Zurich. 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. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes ETH Zurich. 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. Universities had a net outflow of 2,068 moves in AI careersIncludes ETH Zurich. Our estimates of career flows and median h-index across universities and companies in AI, including the role of joint appointments and earlier academic jobs.

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 Valentina Pyatkin yet. Send us a free-licensed photo.

  1. Valentina Pyatkin valentinapy.github.io. Checked 20 October 2026.
  2. Asking It All: Generating Contextualized Questions for any Semantic Role aclanthology.org. Checked 20 October 2026.
  3. Valentina Pyatkin - Schmidt Sciences schmidtsciences.org. Checked 20 October 2026.
  4. Draw Me a Flower: Processing and Grounding Abstraction in Natural Language aclanthology.org. Checked 20 October 2026.
  5. OLMo: Accelerating the Science of Language Models aclanthology.org. Checked 20 October 2026.
  6. Best Paper Awards - ACL 2024 2024.aclweb.org. Checked 20 October 2026.
  7. Tülu 3: Pushing Frontiers in Open Language Model Post-Training arxiv.org. Checked 20 October 2026.
  8. Generalizing Verifiable Instruction Following arxiv.org. Checked 20 October 2026.
  9. Olmo 3 arxiv.org. Checked 20 October 2026.