PeopleofSuperintelligence

Researcher

Shanhuang Chen

Shanhuang Chen is a DeepSeek-AI researcher whose work spans language models and the computing systems used to train them. In 2024, he coauthored DeepSeek LLM, DeepSeek-V2, Fire-Flyer AI-HPC, and DeepSeek-V3, linking model scaling and mixture-of-experts design with large-scale GPU infrastructure. He later coauthored DeepSeek-R1 in 2025 and reports on sparse attention, million-token context, and KV-cache compression through September 2026.

Now at
DeepSeek
Works
8
Citations
2,084
h-index
6

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. January 2024

    Scaling DeepSeek LLM

    In January 2024, Chen coauthors DeepSeek LLM, a project that studies scaling laws and presents 7B and 67B open-source language models. The 67B model surpasses LLaMA-2 70B on benchmarks in code, mathematics, and reasoning.

  2. May 2024

    DeepSeek-V2 mixture-of-experts model

    In May 2024, Chen coauthors DeepSeek-V2, a mixture-of-experts language model pretrained on an 8.1-trillion-token corpus. It uses 21B activated parameters and achieves top-tier performance among open-source models.

  3. August 2024

    Fire-Flyer AI-HPC design

    In August 2024, Chen coauthors Fire-Flyer AI-HPC, a hardware-software co-design for deep learning. Fire-Flyer 2 runs on 10,000 PCIe A100 GPUs and approaches DGX-A100 performance while cutting costs by half and energy consumption by 40%.

  4. December 2024

    Research and engineering on V3

    In December 2024, the DeepSeek-V3 report credits Chen under Research & Engineering. The mixture-of-experts model has 671B parameters, activates 37B for each token, and takes 2.788 million H800 GPU hours to train.

  5. September 2025

    DeepSeek-R1 in Nature

    In September 2025, Chen coauthors Nature’s paper “DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learning.” He is affiliated with the DeepSeek-AI Team in Hangzhou.

  6. December 2025

    Sparse attention in V3.2

    In December 2025, Chen coauthors DeepSeek-V3.2, which introduces DeepSeek Sparse Attention to reduce computation in long-context scenarios. Its high-compute variant, DeepSeek-V3.2-Speciale, achieves gold-medal performance at the 2025 International Mathematical Olympiad and International Olympiad in Informatics.

  7. April 2026

    A million-token V4 preview

    In April 2026, Chen coauthors a preview report that introduces DeepSeek-V4-Pro with 1.6T parameters and DeepSeek-V4-Flash with 284B parameters. Both models support a context length of one million tokens.

  8. As of October 2026

    DeepSeek-AI and V4.1-Flash

    Chen works with DeepSeek-AI and coauthors its September 2026 report on DeepSeek-V4.1-Flash, a multimodal mixture-of-experts model with 552B backbone parameters and up to one million tokens of context. It activates 16B parameters per token during decode and 8B during prefill.

Most-cited works

  1. DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learning2025 · Nature · 1,118 citationsWith Bingxuan Wang, Chengqi Deng, Damai Dai, Daya Guo, Erhang Li and Fangyun Lin
  2. Can Open Large Language Models Catch Vulnerabilities?2025 · 539 citationsWith Bingxuan Wang, Chengqi Deng, Damai Dai, Erhang Li, Fangyun Lin and Haocheng Wang
  3. DeepSeek-V3 Technical Report2024 · 292 citationsWith Bingxuan Wang, Chengqi Deng, Damai Dai, Daya Guo, Dejian Yang and Erhang Li
  4. DeepSeek LLM: Scaling Open-Source Language Models with Longtermism2024 · 98 citationsWith Bingxuan Wang, Chengqi Deng, Damai Dai, Daya Guo, Erhang Li and Fangyun Lin
  5. Fire-Flyer AI-HPC: A Cost-Effective Software-Hardware Co-Design for Deep Learning2024 · 2012 International Conference for High Performance Computing, Networking, Storage and Analysis · 21 citationsWith Chengqi Deng, Mingchuan Zhang, Shanghao Lu and Shangyan Zhou

From OpenAlex, as of 21 October 2026.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Shanhuang Chen. 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.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

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

  1. [2401.02954] DeepSeek LLM: Scaling Open-Source Language Models with Longtermism arxiv.org. Checked 21 October 2026.
  2. Paper page - DeepSeek LLM: Scaling Open-Source Language Models with Longtermism huggingface.co. Checked 21 October 2026.
  3. [2405.04434] DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model arxiv.org. Checked 21 October 2026.
  4. [2408.14158] Fire-Flyer AI-HPC: A Cost-Effective Software-Hardware Co-Design for Deep Learning arxiv.org. Checked 21 October 2026.
  5. [2412.19437] DeepSeek-V3 Technical Report arxiv.org. Checked 21 October 2026.
  6. DeepSeek-V3 Technical Report arxiv.org. Checked 21 October 2026.
  7. DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learning | Nature nature.com. Checked 21 October 2026.
  8. [2512.02556] DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models arxiv.org. Checked 21 October 2026.
  9. [2606.19348] DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence arxiv.org. Checked 21 October 2026.
  10. [2609.19969] DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression arxiv.org. Checked 21 October 2026.