PeopleofSuperintelligence

Researcher

Ruochen Wang

Also known as 王若宸

Ruochen Wang is a Research Scientist at OpenAI whose graduate work on neural architecture search earned an ICLR 2021 Outstanding Paper Award. He studied computer science and statistics at the University of Michigan, graduating in 2019, then completed a computer science PhD at UCLA from 2020 to 2024, working with professor Cho-Jui Hsieh. His first-author paper "Rethinking Architecture Selection in Differentiable NAS" received the award in 2021, and his later papers turned to prompt design for language models, including Mixture-of-Prompts at ICML 2024. In March 2025, he coauthored a replication of the "aha moment" of DeepSeek R1 for multimodal reasoning, work from TurningPoint AI Research, the team he founded during his PhD. He is a Research Scientist at Sora, now WorldSim and Robotics, at OpenAI, which he joined in February 2025.

Now at
OpenAI
Badges

Timeline

  1. 2013–2019

    Computer science at Michigan

    He starts in the finance honors program at Shanghai University of Finance and Economics in 2013, then transfers and earns a B.S. in computer science and statistics at the University of Michigan, Ann Arbor, in 2019.

  2. 2020–2024

    A PhD at UCLA

    He pursues a PhD in computer science at the University of California, Los Angeles, from January 2020 to December 2024. His papers during these years, mostly written with UCLA professor Cho-Jui Hsieh, cover neural architecture search, dataset condensation and prompt design.

  3. 2021

    An ICLR Outstanding Paper

    He is first author of "Rethinking Architecture Selection in Differentiable NAS" with Minhao Cheng, Xiangning Chen, Xiaocheng Tang and Cho-Jui Hsieh. The paper finds that the size of an architecture parameter is an unreliable guide to how much an operation contributes to the supernet, proposes a perturbation-based selection rule, and receives an ICLR 2021 Outstanding Paper Award, one of eight among 860 accepted papers.

  4. 2022

    UCLA Outstanding Graduate Student

    UCLA's computer science department names him its Outstanding Graduate Student in May 2022; the department gives the award to one student.

  5. 2024

    Mixture-of-Prompts at ICML

    He is first author of "One Prompt is not Enough: Automated Construction of a Mixture-of-Expert Prompts" at ICML 2024, whose Mixture-of-Prompts method reaches an average win rate of 81% against prior methods across several major benchmarks. In June 2024, he also posts "Large Language Models are Interpretable Learners" with Si Si, Felix Yu, Dorothea Wiesmann, Cho-Jui Hsieh and Inderjit Dhillon, which combines a pretrained LLM with symbolic programs and appears at ICLR 2025.

  6. March 2025

    Aha moment in visual reasoning

    In March 2025, he coauthors "R1-Zero's 'Aha Moment' in Visual Reasoning on a 2B Non-SFT Model" with Hengguang Zhou, Xirui Li, Minhao Cheng, Tianyi Zhou and Cho-Jui Hsieh, the first successful replication of R1's emergent behavior for multimodal reasoning. The work comes from TurningPoint AI Research, which he founded and led during his PhD as a team of 10 students and professors across several labs.

  7. As of October 2026

    Research Scientist at OpenAI

    He joins OpenAI in February 2025 after his PhD, where he trains and demos ChatGPT-Voice on the GPT-5 livestream. He is a Research Scientist at Sora, now WorldSim and Robotics, on full-stack research.

Career

  1. Since 2025OpenAIResearch Scientist
  2. TurningPoint AI ResearchFounder

Education

  1. 2024University of California, Los AngelesPhD
  2. 2019University of MichiganB.S.
  3. From 2013Shanghai University of Finance and Economics

From the biography.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Ruochen Wang. 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. 16% of AI career moves in the 2020s went from a university to a companyIncludes Ruochen Wang. 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.
  3. Among employers with 30+ full-time cases, OpenAI leads primary technical role H-1B pay at $310,000, and Amazon, Google and Microsoft file 57% of all casesIncludes OpenAI. Across 440,669 certified US H-1B filings from 44 employers (FY2020–FY2026), OpenAI leads primary technical role base pay at $310,000 and xAI leads overall base pay at $290,000.
  4. 22 of 133 people shaping AI with a former OpenAI role now work at AnthropicIncludes OpenAI. Our estimates of where OpenAI recruits came from and where 133 people with former OpenAI roles work now, including the companies and labs they founded.

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 Ruochen Wang yet. Send us a free-licensed photo.

  1. Ruochen Wang 王若宸 ruocwang.github.io. Checked 22 October 2026.
  2. Professor Hsieh and His Students Won the ICLR 2021 Outstanding Paper Award | CS cs.ucla.edu. Checked 22 October 2026.
  3. [2108.04392] Rethinking Architecture Selection in Differentiable NAS arxiv.org. Checked 22 October 2026.
  4. [2407.00256] One Prompt is not Enough: Automated Construction of a Mixture-of-Expert Prompts arxiv.org. Checked 22 October 2026.
  5. [2406.17224] Large Language Models are Interpretable Learners arxiv.org. Checked 22 October 2026.
  6. Large Language Models Are Interpretable Learners | ML Anthology mlanthology.org. Checked 22 October 2026.
  7. [2503.05132] R1-Zero's "Aha Moment" in Visual Reasoning on a 2B Non-SFT Model arxiv.org. Checked 22 October 2026.