PeopleofSuperintelligence

Researcher

Zizheng Pan

Also known as 潘梓正

Zizheng Pan is a computer scientist whose work spans efficient vision transformers and multimodal AI. He studied software engineering at Harbin Institute of Technology, Weihai, and computer science at the University of Adelaide before completing a PhD at Monash University in 2024. After a 2023 research internship at NVIDIA, he contributed to DeepSeek-VL2 and DeepSeek-V3 in 2024, then coauthored DeepSeek-R1 and Janus-Pro in 2025; he also coauthored NVIDIA's 2025 T-Stitch paper. He is a full-time researcher at DeepSeek, where he works on multimodal LLMs.

Now at
DeepSeek
Badges

Timeline

  1. 2015–2019

    Software engineering at Harbin

    He studies software engineering at Harbin Institute of Technology, Weihai, from 2015 to 2019.

  2. 2018–2020

    Computer science at Adelaide

    He studies computer science at the University of Adelaide from 2018 to 2020.

  3. 2021–2024

    A Monash PhD in visual transformers

    He completes a computer-science PhD at Monash University in 2024. In 2021, he coauthors “Scalable Visual Transformers with Hierarchical Pooling,” which progressively pools visual tokens to reduce sequence length and computational cost.

  4. July–October 2023

    NVIDIA research internship

    He serves as a research intern in NVIDIA's AI Algorithm Group from July to October 2023.

  5. December 2024

    DeepSeek-VL2 and V3 contributions

    In December 2024, the DeepSeek-VL2 paper names him as a core contributor to a mixture-of-experts vision-language model. The DeepSeek-V3 technical report lists him in its Research & Engineering credits.

  6. January 2025

    DeepSeek-R1 and Janus-Pro

    In January 2025, he coauthors DeepSeek-R1, which uses reinforcement learning to strengthen language-model reasoning, and Janus-Pro, which combines multimodal understanding with text-to-image generation.

  7. April 2025

    Sampling faster with T-Stitch

    At ICLR in April 2025, he coauthors T-Stitch, a method that accelerates sampling in pretrained diffusion models.

  8. Since July 2024

    Researcher at DeepSeek

    He joins DeepSeek as a full-time researcher in July 2024 and works on multimodal LLMs there in 2026.

Career

  1. Since 2024DeepSeekResearcher
  2. 2023NVIDIAResearch intern

Education

  1. 2024Monash UniversityPhD
  2. 2020University of Adelaide
  3. 2019Harbin Institute of Technology (Weihai)

From the biography.

  1. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Zizheng Pan. 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.
  2. 16% of AI career moves in the 2020s went from a university to a companyIncludes Zizheng Pan. 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. 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.
  4. 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.

Colleagues

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

Sources and credits

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

  1. Zizheng Pan zizhengpan.github.io. Checked 20 October 2026.
  2. How an Australian uni student helped create China's revolutionary AI platform DeepSeek abc.net.au. Checked 20 October 2026.
  3. Scalable vision transformers with hierarchical pooling research.monash.edu. Checked 20 October 2026.
  4. DeepSeek-VL2: Mixture-of-Experts Vision-Language Models for Advanced Multimodal Understanding arxiv.org. Checked 20 October 2026.
  5. DeepSeek-V3 Technical Report arxiv.org. Checked 20 October 2026.
  6. DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning arxiv.org. Checked 20 October 2026.
  7. Janus-Pro: Unified Multimodal Understanding and Generation with Data and Model Scaling arxiv.org. Checked 20 October 2026.
  8. T-Stitch: Accelerating Sampling in Pre-Trained Diffusion Models with Trajectory Stitching | NVIDIA Learning and Perception Research research.nvidia.com. Checked 20 October 2026.