PeopleofSuperintelligence

Researcher

Xiao Bi

Xiao Bi is a core researcher at DeepSeek whose work spans language models, mathematical reasoning and AI training systems. He coauthored DeepSeek LLM and DeepSeek-Coder in January 2024, then DeepSeekMath in February 2024 and DeepSeek-V2 in May 2024. At DeepSeek AI, he coauthored Fire-Flyer AI-HPC in August 2024 and DeepSeek-V3 in December 2024, then DeepSeek-R1 in January 2025. In September 2026, he coauthored DeepSeek Elastic Compute, a production sandbox platform for agent training.

Now at
DeepSeek
Works
3
Citations
24
h-index
2

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. January 2024

    DeepSeek LLM and Coder

    In January 2024, he coauthors DeepSeek LLM, a project for open-source language models, and DeepSeek-Coder, an open-source code-model series trained from scratch on 2 trillion tokens.

  2. February 2024

    DeepSeekMath training

    In February 2024, he coauthors DeepSeekMath, which continues pretraining DeepSeek-Coder-Base-v1.5 7B on 120 billion math-related tokens. The 7B model scores 51.7% on MATH without external toolkits or voting.

  3. May 2024

    DeepSeek-V2 efficiency

    In May 2024, he coauthors DeepSeek-V2, a mixture-of-experts model that saves 42.5% of training costs compared with DeepSeek 67B. It cuts KV cache by 93.3% and raises maximum generation throughput to 5.76 times.

  4. August 2024

    Fire-Flyer AI-HPC

    At DeepSeek AI, he coauthors Fire-Flyer AI-HPC in August 2024, a software-hardware design for deep-learning infrastructure. The Fire-Flyer 2 setup uses 10,000 PCIe A100 GPUs, reaches performance near DGX-A100, and reduces costs by half and energy use by 40%.

  5. December 2024

    DeepSeek-V3 technical report

    In December 2024, he coauthors DeepSeek-V3, a mixture-of-experts model with 671 billion total parameters and 37 billion active per token. Its training uses 14.8 trillion tokens.

  6. January 2025

    DeepSeek-R1 reasoning

    In January 2025, he coauthors DeepSeek-R1, whose large-scale reinforcement-learning framework develops reasoning patterns such as self-reflection, verification and dynamic strategy adaptation.

  7. September 2026

    DeepSeek Elastic Compute

    In September 2026, he coauthors DeepSeek Elastic Compute (DSec), a production sandbox platform for agent training. It offers FnCall, container, microVM and full-VM backends through one SDK, and coordinates placement and lifecycle management across a cluster.

  8. As of October 2026

    Core researcher at DeepSeek

    As of October 2026, he is a core researcher at DeepSeek, with work on DeepSeek LLM, DeepSeek-V2, DeepSeek-V3, DeepSeek-R1 and DeepSeekMath.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Xiao Bi. 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 Xiao Bi yet. Send us a free-licensed photo.

  1. DeepSeek LLM: Scaling Open-Source Language Models with Longtermism arxiv.org. Checked 22 October 2026.
  2. DeepSeek-Coder: When the Large Language Model Meets Programming -- The Rise of Code Intelligence arxiv.org. Checked 22 October 2026.
  3. DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models arxiv.org. Checked 22 October 2026.
  4. DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model arxiv.org. Checked 22 October 2026.
  5. Fire-Flyer AI-HPC: A Cost-Effective Software-Hardware Co-Design for Deep Learning arxiv.org. Checked 22 October 2026.
  6. Contributors - SC24 Schedule sc24.conference-program.com. Checked 22 October 2026.
  7. DeepSeek-V3 Technical Report arxiv.org. Checked 22 October 2026.
  8. DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning arxiv.org. Checked 22 October 2026.
  9. DeepSeek Elastic Compute (DSec): A Sandbox Infrastructure for Effective Agentic Training at Scale arxiv.org. Checked 22 October 2026.
  10. DeepSeek | Lab Index labindex.ai. Checked 22 October 2026.