PeopleofSuperintelligence

Researcher

Yuheng Zou

Yuheng Zou is an AI researcher and software engineer who contributed to the DeepSeek model series, from DeepSeek-V2 to DeepSeek-V4.1-Flash. In 2016, they coauthored DoReFa-Net, a method for training convolutional neural networks with low bitwidth weights, activations and gradients. Their name appears on the DeepSeek-V2 paper in May 2024, the Fire-Flyer AI-HPC paper in August 2024, DeepSeek-V3 in December 2024, DeepSeek-R1 in January 2025 and DeepSeek-V3.2 in December 2025. Nature published the DeepSeek-R1 paper in September 2025, and DeepSeek-V4 followed in April 2026 and DeepSeek-V4.1-Flash in September 2026. As of October 2026, they are a researcher at Peking University and a software engineer at 幻方AI.

Timeline

  1. 2016–2017

    DoReFa-Net and quantization

    In 2016, they coauthor DoReFa-Net with Shuchang Zhou, Yuxin Wu, Zekun Ni, Xinyu Zhou and He Wen, a method for training convolutional neural networks with low bitwidth weights, activations and gradients. In 2017, they are an author of Balanced Quantization, published in the Journal of Computer Science and Technology.

  2. May 2024

    DeepSeek-V2 language model

    In May 2024, they are among the authors of DeepSeek-V2, a Mixture-of-Experts language model with 236B total parameters, of which 21B are activated for each token. It supports a context length of 128K tokens.

  3. August 2024

    Fire-Flyer AI-HPC co-design

    In August 2024, they are among the authors of Fire-Flyer AI-HPC, a hardware-software co-design framework for deep learning. The paper describes the Fire-Flyer 2 cluster, deployed with 10,000 PCIe A100 GPUs.

  4. December 2024

    DeepSeek-V3 technical report

    In December 2024, they are among the authors of the DeepSeek-V3 Technical Report. DeepSeek-V3 is a Mixture-of-Experts language model with 671B total parameters, of which 37B are activated for each token.

  5. January 2025–September 2025

    DeepSeek-R1 in Nature

    In January 2025, they are among the authors of DeepSeek-R1, which shows that the reasoning abilities of LLMs can be incentivized through pure reinforcement learning. Nature publishes the paper in September 2025.

  6. December 2025

    DeepSeek-V3.2 sparse attention

    In December 2025, they are among the authors of DeepSeek-V3.2. The model introduces DeepSeek Sparse Attention, an attention mechanism that reduces computational complexity in long-context scenarios.

  7. April 2026–September 2026

    DeepSeek-V4 and V4.1-Flash

    In April 2026, they are among the authors of the DeepSeek-V4 preview, two Mixture-of-Experts models that support a context length of one million tokens. In September 2026, they are among the authors of DeepSeek-V4.1-Flash, a multimodal Mixture-of-Experts model with 552B backbone parameters.

  8. As of October 2026

    Peking University and 幻方AI

    As of October 2026, they are a researcher at Peking University and a software engineer at 幻方AI.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Peking University. 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 company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Peking University. 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.
  3. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes Peking University. 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. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Peking University. 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.

Sources and credits

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

  1. Yuheng Zou - Google Scholar scholar.google.com.tw. Checked 21 October 2026.
  2. DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients arxiv.org. Checked 21 October 2026.
  3. DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model arxiv.org. Checked 21 October 2026.
  4. Yuheng Zou | alphaXiv alphaxiv.org. Checked 21 October 2026.
  5. Fire-Flyer AI-HPC: A Cost-Effective Software-Hardware Co-Design for Deep Learning arxiv.org. Checked 21 October 2026.
  6. DeepSeek-V3 Technical Report arxiv.org. Checked 21 October 2026.
  7. DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning arxiv.org. Checked 21 October 2026.
  8. DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learning nature.com. Checked 21 October 2026.
  9. DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models arxiv.org. Checked 21 October 2026.
  10. DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence arxiv.org. Checked 21 October 2026.
  11. DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression arxiv.org. Checked 21 October 2026.