PeopleofSuperintelligence

Researcher

Peiyi Wang

Also known as 王培懿

Peiyi Wang is an NLP researcher whose work spans few-shot event classification, document-level event extraction and language-model reasoning. He coauthored a 2021 study of trigger bias, the TSAR event-extraction model in 2022, and Math-Shepherd in 2023. In 2024, he coauthored DeepSeekMath, which introduced GRPO, and the DeepSeek-V3 report; he coauthored DeepSeek-R1 in 2025 and the DeepSeek-V4 preview in April 2026. As of October 2026, he is a researcher at DeepSeek.

Now at
DeepSeek
Works
60
Badges

Publication numbers from OpenAlex, as of 22 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.

Timeline

  1. 2021

    Behind the scenes

    In 2021, he coauthors Behind the Scenes, which identifies trigger biases in few-shot event classification and proposes two evaluation sampling methods.

  2. 2022

    Two-stream event extraction

    In 2022, he coauthors TSAR, a two-stream model that uses abstract meaning representation graphs to extract event arguments across full documents.

  3. 2023

    Math-Shepherd process rewards

    In 2023, he coauthors Math-Shepherd, a process reward model trained on automatically constructed process-wise supervision data for language-model reasoning.

  4. 2024

    DeepSeekMath and GRPO

    In 2024, he coauthors DeepSeekMath, which introduces Group Relative Policy Optimization (GRPO) to strengthen mathematical reasoning.

  5. 2024

    The DeepSeek-V3 report

    In 2024, he coauthors the DeepSeek-V3 technical report, which describes a mixture-of-experts model with 671 billion total parameters and 37 billion active for each token.

  6. 2025

    DeepSeek-R1 reasoning

    In 2025, he coauthors DeepSeek-R1, which shows that reinforcement learning can encourage language-model reasoning without human-labeled reasoning trajectories.

  7. April 2026

    The DeepSeek-V4 preview

    In April 2026, he coauthors the DeepSeek-V4 preview, which introduces two mixture-of-experts language models.

  8. As of October 2026

    Researcher at DeepSeek

    As of October 2026, he is a researcher at DeepSeek. His report work includes DeepSeek LLM, DeepSeekMath, DeepSeek-V3 and DeepSeek-R1.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Peiyi 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 Peiyi 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. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes DeepSeek. Our estimates of the organizations where 550 people worked directly before joining 9 frontier AI labs across 620 career transitions, comparing the main feeder organizations per lab.
  4. 5,000 frontier users would be 38.2% of the core labs' 13,100 reported staffIncludes DeepSeek. 5,000 frontier users would be 38.2% of the core labs' 13,100 reported staff. We test the estimate against 3,913 named model contributors and dated all-staff records across Google DeepMind, OpenAI and Anthropic, plus six additional frontier-grade labs.

Sources and credits

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

  1. Behind the Scenes: An Exploration of Trigger Biases Problem in Few-Shot Event Classification arxiv.org. Checked 21 October 2026.
  2. A Two-Stream AMR-enhanced Model for Document-level Event Argument Extraction aclanthology.org. Checked 21 October 2026.
  3. Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotations arxiv.org. Checked 21 October 2026.
  4. DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models arxiv.org. Checked 21 October 2026.
  5. DeepSeek-V3 Technical Report arxiv.org. Checked 21 October 2026.
  6. DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning arxiv.org. Checked 21 October 2026.
  7. DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence arxiv.org. Checked 21 October 2026.
  8. Peiyi Wang | Researcher at DeepSeek | LLMpeople llm-people.com. Checked 21 October 2026.