PeopleofSuperintelligence

Researcher

Deli Chen

Also known as 陈德里

Deli Chen is an AI researcher whose work spans graph neural networks, language models, and reasoning systems. He earned degrees in information management and computer science at Peking University, then researched natural language processing and language models at Tencent WeChat AI. At DeepSeek, he coauthored studies of in-context learning, mixture-of-experts models, and reinforcement-learning-based reasoning. He is a senior researcher at DeepSeek and, in May 2026, called for hires to build Code Harness.

Now at
DeepSeek
Works
12
Badges

Publication numbers from OpenAlex, as of 19 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.

Timeline

  1. 2015–2019

    A bachelor's at Peking University

    He earns a B.S. in information management at Peking University's School of Information Management in 2019.

  2. 2019–2021

    Computer science at Peking University

    He earns an M.S. in computer science at Peking University in 2021, working in its MOE Key Lab of Computational Linguistics under adviser Xu Sun.

  3. 2020

    The over-smoothing problem

    In 2020, he coauthors a paper that measures and relieves over-smoothing in graph neural networks from a topological view.

  4. 2021–2023

    Research at Tencent WeChat AI

    He works as a researcher at Tencent WeChat AI from 2021 to 2023, focusing on NLP and language models.

  5. 2023

    Label Words are Anchors

    At EMNLP 2023, he coauthors Label Words are Anchors, a study of how example labels guide in-context learning.

  6. 2024

    DeepSeekMoE and V3

    In 2024, he coauthors DeepSeekMoE, a paper on expert specialization in mixture-of-experts language models. The DeepSeek-V3 technical report also lists him as an author and says the model used 14.8 trillion training tokens.

  7. 2025

    DeepSeek-R1 and AI's impact

    In 2025, he coauthors the DeepSeek-R1 paper on reinforcement learning for language-model reasoning. In November 2025, he says technology companies should protect human safety and help shape societal order.

  8. Since 2023

    Senior researcher at DeepSeek

    He joins DeepSeek as a senior researcher in 2023 and contributes to its models. In May 2026, he calls for hires to build Code Harness from scratch, with the goal of benchmarking against Claude Code.

Career

  1. Since 2023DeepSeekSenior Researcher
  2. 2021–2023Tencent Holdings LimitedResearcher
  3. 2019–2021Peking University

Education

  1. 2021Peking UniversityM.S. in Computer Science
  2. 2019Peking UniversityB.S. in Information Management

From the biography.

  1. The median h-index is 53 at universities and 19 at frontier AI labsIncludes DeepSeek. The median h-index is 53 at universities and 19 at frontier AI labs. We compare publishing careers, citations and moves to estimate how universities and frontier labs contribute to AI research.
  2. The company share of AI researchers’ top papers rose from 2% to 32%Includes DeepSeek. We estimate where leading AI researchers worked when their most-cited papers appeared. The company share rose from 2% in 1975–1979 to 32% in 2020–2024.
  3. Google has the highest combined h-index among AI organizationsIncludes DeepSeek. Our estimates of combined and median h-index at AI organizations, using current affiliations and OpenAlex records.
  4. Google alumni founded or led 87 other organizations, the most among the labs we compareIncludes DeepSeek. We trace 212 organizations founded or led by AI lab alumni, and 223 career moves between the labs.

Colleagues

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

Sources and credits

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

  1. Deli Chen - DeepSeek AI Researcher victorchen96.github.io. Checked 19 October 2026.
  2. Vol. 34 No. 04: AAAI-20 Technical Tracks 4 ojs.aaai.org. Checked 19 October 2026.
  3. Label Words are Anchors: An Information Flow Perspective for Understanding In-Context Learning aclanthology.org. Checked 19 October 2026.
  4. DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models aclanthology.org. Checked 19 October 2026.
  5. DeepSeek-V3 Technical Report arxiv.org. Checked 19 October 2026.
  6. DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning arxiv.org. Checked 19 October 2026.
  7. An AI executive's dire warnings about the future are chilling – but his solution is worse than the problem techradar.com. Checked 19 October 2026.
  8. China’s DeepSeek Starts Hiring for Agent Harness Team, Taking Aim at Anthropic's Claude Code yicaiglobal.com. Checked 19 October 2026.