PeopleofSuperintelligence

Researcher

Keqin Bao

Also known as 鲍克勤

Keqin Bao is a language model researcher on Alibaba's Qwen team in Hangzhou, and his doctoral work applied large language models to recommender systems. After a computer science degree at Harbin Institute of Technology, he began a PhD in 2021 at the USTC Lab for Data Science at the University of Science and Technology of China. Early work on translation evaluation produced Alibaba-Translate China's entry to the WMT 2022 quality estimation task, which ranked first overall in two settings. As first author, he introduced TALLRec in 2023, a framework for tuning language models with recommendation data, and Decoding Matters at EMNLP 2024, a new decoding method for LLM-based recommenders. He coauthored the Qwen2.5 and Qwen3 technical reports in 2024 and 2025, then completed his PhD in 2026 and joined Qwen.

Now at
Alibaba
Badges

Timeline

  1. 2017–2021

    Computer science at HIT

    He studies computer science at Harbin Institute of Technology from 2017 to 2021 and earns a BEng.

  2. 2021

    Doctoral studies at USTC

    In 2021, he begins a PhD in cyberspace security at the University of Science and Technology of China, working in the USTC Lab for Data Science within the School of Cyber Science and Technology. His research moves from translation evaluation to large language models for recommendation.

  3. 2022

    Translation quality estimation at WMT

    In 2022, he is first author of Alibaba-Translate China's submission to the WMT 2022 Quality Estimation Shared Task, built on the UniTE (Unified Translation Evaluation) framework. The system ranks first overall in the multilingual and English-Russian settings.

  4. 2023

    TALLRec for recommendation

    He is first author of TALLRec, a framework that tunes large language models with recommendation data, published at the ACM Conference on Recommender Systems (RecSys) in 2023. TALLRec improves recommendation in the movie and book domains with fewer than 100 training samples and runs on a single RTX 3090 with LLaMA-7B.

  5. 2024

    Decoding Matters at EMNLP

    At EMNLP 2024, he is first author of Decoding Matters, which introduces Debiasing-Diversifying Decoding (D3) for LLM-based recommendation. The method counters amplification bias from ghost tokens and the tendency to generate similar or repetitive items for a user.

  6. December 2024

    The Qwen2.5 technical report

    In December 2024, he coauthors the Qwen2.5 Technical Report, which describes a series of large language models whose pre-training data grows from 7 trillion to 18 trillion tokens. Hosted Qwen2.5-Turbo and Qwen2.5-Plus models are offered through Alibaba Cloud Model Studio.

  7. May 2025

    The Qwen3 technical report

    In May 2025, he coauthors the Qwen3 Technical Report. Qwen3 joins thinking and non-thinking modes in one framework, ranges from 0.6 to 235 billion parameters, supports 119 languages and dialects, and is released under Apache 2.0.

  8. Since 2026

    Joining Alibaba's Qwen team

    He completes his PhD at the USTC Lab for Data Science in 2026 and moves to Alibaba's Qwen team in Hangzhou. In 2026, he also coauthors MTR-Bench, a 3,600-instance benchmark for multi-turn reasoning, at the 64th Annual Meeting of the Association for Computational Linguistics.

Career

  1. Since 2026AlibabaLanguage model researcher

Education

  1. 2021Harbin Institute of TechnologyBEng
  2. From 2021University of Science and Technology of ChinaPhD

From the biography.

  1. 16% of AI career moves in the 2020s went from a university to a companyIncludes Keqin Bao. 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.
  2. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Alibaba. 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.
  3. Universities (31%) and startups and other companies (25%) employ the largest shares of 2020–2026 entrantsIncludes Alibaba. Among 819 people whose first recorded role started in 2020–2026 and who have a known current organization, 60% work in commercial sectors (25% at startups and other companies, 23% at frontier labs and 13% at big tech, with 9 in two), and 31% work at universities, compared with 66% for earlier entrants.
  4. The top 10 organizations employed 22% of people shaping AI in 2025: 51% of company researchers and 21% of university researchersIncludes Alibaba. We estimate the share of people shaping AI who worked at the top 10 organizations each year from 1955 to 2025, comparing concentration across companies, universities, and institutes and labs.

Colleagues

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

Sources and credits

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

  1. Keqin Bao | USTC Lab for Data Science data-science.ustc.edu.cn. Checked 20 October 2026.
  2. Keqin Bao - ACL Anthology aclanthology.org. Checked 20 October 2026.
  3. [2305.00447] TALLRec: An Effective and Efficient Tuning Framework to Align Large Language Model with Recommendation arxiv.org. Checked 20 October 2026.
  4. [2412.15115] Qwen2.5 Technical Report arxiv.org. Checked 20 October 2026.
  5. [2505.09388] Qwen3 Technical Report arxiv.org. Checked 20 October 2026.
  6. People | USTC Lab for Data Science data-science.ustc.edu.cn. Checked 20 October 2026.