PeopleofSuperintelligence

Researcher

Tianhao Li

Tianhao Li is an AI researcher whose work connects multilingual language models with the Qwen family and machine translation. He coauthored PolyLM in 2023 and MoE-CT in 2024, which explored multilingual training and expansion. He coauthored Qwen2 and Qwen2.5 in 2024, Qwen3 in 2025, and Qwen-Scope for mechanistic interpretability in 2026. In October 2026, he coauthored Qwen-MT-Lumen, an instruction-specialized translation system using Qwen3-14B across 23 directions, and is affiliated with Alibaba Token Hub.

Now at
Alibaba
Works
35
Citations
642
h-index
11
Badges

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. July 2023

    PolyLM multilingual model

    In July 2023, he coauthors PolyLM, a multilingual language model trained on 640 billion tokens and available in 1.7B and 13B sizes.

  2. June 2024

    MoE-CT multilingual training

    In June 2024, he coauthors MoE-CT, an approach for continual large-language-model training that separates base-model learning from multilingual expansion.

  3. July 2024

    Qwen2 technical report

    In July 2024, he coauthors the Qwen2 Technical Report, which presents foundational and instruction-tuned models from 0.5 to 72 billion parameters, including dense models and a Mixture-of-Experts model.

  4. December 2024

    Qwen2.5 technical report

    In December 2024, he coauthors the Qwen2.5 Technical Report, which scales high-quality pretraining data from 7 trillion to 18 trillion tokens. The Qwen2.5 family also includes an instruction-tuned 1.5B model.

  5. May 2025

    Qwen3 technical report

    In May 2025, he coauthors the Qwen3 Technical Report, which expands multilingual support from 29 to 119 languages and dialects.

  6. May 2026

    Qwen-Scope sparse autoencoders

    In May 2026, he coauthors Qwen-Scope, an open-source suite of sparse autoencoders for mechanistic interpretability with 14 groups across seven Qwen3 and Qwen3.5 model variants.

  7. October 2026

    Qwen-MT-Lumen for WMT

    In October 2026, he coauthors Qwen-MT-Lumen, an instruction-specialized translation system built on a Qwen3-14B backbone for WMT 2026's 23 directions.

  8. As of October 2026

    Alibaba Token Hub affiliation

    As of October 2026, he is affiliated with Alibaba Token Hub.

Career

  1. NowAlibaba
  2. 2023–2026China University of Petroleum, East China

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

  1. 16% of AI career moves in the 2020s went from a university to a companyIncludes Tianhao Li. 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 Tianhao Li yet. Send us a free-licensed photo.

  1. PolyLM: An Open Source Polyglot Large Language Model arxiv.org. Checked 22 October 2026.
  2. MoE-CT: A Novel Approach For Large Language Models Training With Resistance To Catastrophic Forgetting arxiv.org. Checked 22 October 2026.
  3. Qwen2 Technical Report arxiv.org. Checked 22 October 2026.
  4. Hello Qwen2 | Qwen qwenlm.github.io. Checked 22 October 2026.
  5. Qwen2.5 Technical Report arxiv.org. Checked 22 October 2026.
  6. Qwen/Qwen2.5-1.5B-Instruct · Hugging Face huggingface.co. Checked 22 October 2026.
  7. Qwen3 Technical Report arxiv.org. Checked 22 October 2026.
  8. Qwen-Scope: Turning Sparse Features into Development Tools for Large Language Models arxiv.org. Checked 22 October 2026.
  9. Qwen-MT-Lumen at WMT2026: Instruction-Specialized Translation via Quality-Aware Training and On-Policy Distillation arxiv.org. Checked 22 October 2026.
  10. Qwen-MT-Lumen at WMT2026: Instruction-Specialized Translation via Quality-Aware Training and On-Policy Distillation arxiv.org. Checked 22 October 2026.