PeopleofSuperintelligence

Researcher

Yang Zhang

Yang Zhang is a DeepSeek-AI research contributor whose work spans language models, reinforcement learning and long-context systems. In 2024, he coauthored DeepSeek-V3, a 671-billion-parameter mixture-of-experts model with 37 billion active per token. In 2025, he coauthored a DeepSeek-R1 paper on reasoning through pure reinforcement learning and contributed to DeepSeek-V3.2's sparse-attention and reinforcement-learning work on reasoning and agent performance. In 2026, he contributed to reports on DeepSeek-V4's million-token context and DeepSeek-V4.1-Flash's KV-cache compression. By September 2026, he is a member of DeepSeek-AI's Research & Engineering team.

Now at
DeepSeek

Timeline

  1. 2024

    DeepSeek-V3's 671B model

    In 2024, he contributes to DeepSeek-V3, a mixture-of-experts language model with 671 billion total parameters and 37 billion active per token.

  2. 2025

    Pure reinforcement learning in DeepSeek-R1

    In 2025, he coauthors a DeepSeek-R1 paper showing that pure reinforcement learning can incentivize reasoning in language models without human-labelled reasoning trajectories.

  3. 2025

    Sparse attention in DeepSeek-V3.2

    In 2025, he contributes to DeepSeek-V3.2, whose report introduces DeepSeek Sparse Attention and a scalable reinforcement-learning framework for reasoning and agent performance.

  4. 2026

    Million-token context in DeepSeek-V4

    In 2026, he contributes to the DeepSeek-V4 preview report, which describes two models that support a context length of one million tokens.

  5. September 2026

    KV-cache compression in V4.1-Flash

    In September 2026, he coauthors the DeepSeek-V4.1-Flash report, which combines cross-layer KV-cache reuse with FP4 caching to reduce the model's KV-cache footprint.

  6. By September 2026

    Research and engineering at DeepSeek

    By September 2026, he is a member of DeepSeek-AI's Research & Engineering team.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes DeepSeek. 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. 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.
  3. 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.
  4. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes DeepSeek. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.

Sources and credits

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

  1. DeepSeek-V3 Technical Report arxiv.org. Checked 22 October 2026.
  2. DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learning | Nature nature.com. Checked 22 October 2026.
  3. DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models arxiv.org. Checked 22 October 2026.
  4. DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence arxiv.org. Checked 22 October 2026.
  5. DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression arxiv.org. Checked 22 October 2026.
  6. Yang Zhang | 500AI 500ai.org. Checked 22 October 2026.