PeopleofSuperintelligence

Researcher

Chujie Zheng

Also known as 郑楚杰

Chujie Zheng is a Qwen researcher whose work spans reinforcement learning for language models, Chinese dialogue, and mathematical reasoning evaluation. He studied mathematics, physics, and computer science at Tsinghua University and visited UCLA as a researcher hosted by Nanyun Peng. His work includes EVA, the Yi-Lightning report, ProcessBench, and Group Sequence Policy Optimization (GSPO), which Qwen applied in large-scale training of Qwen3 models; ACL included ProcessBench in its 2025 SAC Highlights. As of October 2026, he is a member of technical staff on the Qwen Team and leads the OPD team for post-training consolidation.

Now at
Qwen Team
Works
49
Citations
1,367
h-index
17
Badges

Publication numbers from OpenAlex, as of 19 October 2026.

Timeline

  1. 2016–2020

    Mathematics and physics at Tsinghua

    From 2016 to 2020, he studies mathematics and physics at Tsinghua University and earns a B.Sc. in both subjects.

  2. 2020–2025

    Doctoral study at Tsinghua

    From 2020 to 2025, he pursues a Ph.D. in Computer Science and Technology at Tsinghua University with Minlie Huang as his adviser.

  3. 2021

    EVA dialogue system

    In 2021, he coauthors EVA, a Chinese dialogue system whose pre-trained model has 2.8 billion parameters.

  4. November 2023–June 2024

    Visiting research at UCLA

    He works as a visiting researcher at UCLA from November 2023 to June 2024, with Nanyun Peng as his host.

  5. 2024

    Yi-Lightning technical report

    In 2024, he coauthors the Yi-Lightning technical report, which presents a large language model that ranks sixth overall on Chatbot Arena.

  6. 2025

    ProcessBench for math reasoning

    In 2025, he coauthors ProcessBench, a 3,400-case benchmark that tests whether models can identify errors in mathematical reasoning. ACL's 2025 SAC Highlights list the paper.

  7. 2025

    Qwen3 and GSPO training

    In 2025, he coauthors the Qwen3 technical report and Group Sequence Policy Optimization (GSPO), a method for reinforcement learning with language models. Qwen applies GSPO to large-scale training of Qwen3 models.

  8. As of October 2026

    Leading Qwen post-training work

    As of October 2026, he is a member of technical staff on the Qwen Team, works on reinforcement learning scaling, and leads the OPD team for post-training consolidation.

Career

  1. NowQwen TeamMember of Technical Staff; OPD Team Lead
  2. 2023–2024University of California, Los AngelesVisiting Researcher
  3. 2020–2023State Key Laboratory of Intelligent Technology and Systems
  4. 2019–2025Beijing National Research Center for Information Science and Technology

Education

  1. 2025Tsinghua UniversityPh.D.
  2. 2020Tsinghua UniversityB.Sc.

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

Most-cited works

  1. Towards Emotional Support Dialog Systems2021 · 226 citations
  2. CEM: Commonsense-Aware Empathetic Response Generation2022 · Proceedings of the AAAI Conference on Artificial Intelligence · 154 citations
  3. Qwen3 Technical Report2025 · 147 citationsWith Baosong Yang, Beichen Zhang, Chengen Huang, Chenxu Lv, Fei Huang and Le Yu
  4. KdConv: A Chinese Multi-domain Dialogue Dataset Towards Multi-turn Knowledge-driven Conversation2020 · Annual Meeting of the Association for Computational Linguistics (ACL) · 104 citations
  5. COLD: A Benchmark for Chinese Offensive Language Detection2022 · Conference on Empirical Methods in Natural Language Processing (EMNLP) · 102 citationsWith Helen Meng

From OpenAlex, as of 19 October 2026.

  1. Stanford University trained or employed 20 of the 89 AI major-prize winners, more than any other universityIncludes University of California, Los Angeles. Stanford University trained or employed 20 of the 89 AI major-prize winners, more than any other university. We compare doctorate and employment links for major-prize and best-paper winners.
  2. Google has the highest combined h-index among AI organizationsIncludes Tsinghua University. Our estimates of combined and median h-index at AI organizations, using current affiliations and OpenAlex records.
  3. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Tsinghua University. Our estimates of where the people shaping AI studied and where their careers led, from education records in cited biographies.
  4. The company share of AI researchers’ top papers rose from 2% to 32%Includes Tsinghua University. 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.

Colleagues

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

Sources and credits

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

  1. Chujie Zheng - Chujie Zheng 郑楚杰 chujiezheng.github.io. Checked 19 October 2026.
  2. EVA: An Open-Domain Chinese Dialogue System with Large-Scale Generative Pre-Training arxiv.org. Checked 19 October 2026.
  3. Yi-Lightning Technical Report arxiv.org. Checked 19 October 2026.
  4. Chujie Zheng - ACL Anthology aclanthology.org. Checked 19 October 2026.
  5. Awards - ACL 2025 2025.aclweb.org. Checked 19 October 2026.
  6. Qwen3 Technical Report arxiv.org. Checked 19 October 2026.
  7. GSPO: Towards Scalable Reinforcement Learning for Language Models | Qwen qwenlm.github.io. Checked 19 October 2026.