AI People Wiki

Builders & open source

Jianfeng Gao

Technical Fellow & Corporate Vice President, Microsoft

Jianfeng Gao is a Technical Fellow and Corporate Vice President at Microsoft. Gao's Microsoft Research page describes Gao setting company-wide direction for the AI models behind Bing, Azure AI, Microsoft 365, and Copilot, and heading a Microsoft Research-backed investment in core capabilities such as multimodal reasoning, agentic coding, tool use, and long-context modeling. Gao was earlier a Distinguished Scientist and Vice President at Microsoft Research and a co-founder of its Deep Learning Group, and co-invented the deep structured semantic models (DSSM) used in web search. Gao is a co-author of the Phi-3 technical report and of LLaVA-Med, a vision-language assistant for biomedicine, and also serves as an Adjunct Professor at the University of Washington.1234

At a glance

Current roles
Technical Fellow & Corporate Vice President +1 more
Sector
Company
Education
UW
Work listed
3
Honors
3
Sources cited
6
Profile reviewed

Career

Jianfeng Gao spent 1999 to 2005 as a research lead in the Natural Language Computing Group at Microsoft Research Asia, where the group built a Chinese speech recognition system shipped with Microsoft Office, Chinese and Japanese input method editors, and the natural language platform for Windows Vista. From 2005 to 2006 Gao was a research lead in Microsoft's Natural Interactive Services Division, working on a natural user interface effort for Windows.1

From 2006 to 2014 Gao was a Principal Researcher in the Natural Language Processing Group at Microsoft Research in Redmond, working on web search, the Bing speller, query understanding, ads prediction and statistical machine translation. Gao later co-founded and long led the Deep Learning Group at Microsoft Research, rising to Distinguished Scientist and Vice President, and now serves as a Technical Fellow and Corporate Vice President.1

Research

Gao's Microsoft Research page credits Gao as a co-inventor of the OWL-QN method, a variant of L-BFGS for optimizing sparse linear models, and of deep structured semantic models (DSSM). The DSSM paper, presented at CIKM in October 2013, learns semantic representations of queries and documents from search clickthrough data.13

With Michel Galley and Lihong Li, Gao wrote "Neural Approaches to Conversational AI", a survey of neural methods for question answering, task-oriented dialogue and chatbots, first posted in 2018. More recent co-authored work includes LLaVA-Med (2023), a vision-language assistant for biomedicine, and the Phi-3 technical report (2024).542

Work and contributions

  • Phi-3 Open weights

    Listed among the authors of the Phi-3 Technical Report (2024), which introduces phi-3-mini, a 3.8 billion parameter model small enough to be deployed on a phone. The phi-3-mini-4k-instruct model card on Hugging Face lists the MIT license.26

  • Deep Structured Semantic Models (DSSM) Published research

    Co-author of "Learning Deep Structured Semantic Models for Web Search using Clickthrough Data" (CIKM 2013); Gao's Microsoft Research page describes Gao as a co-inventor of DSSM.31

  • LLaVA-Med Published research

    Co-author of the 2023 paper "LLaVA-Med: Training a Large Language-and-Vision Assistant for Biomedicine in One Day".4

Timeline

  1. Became a research lead in the Natural Language Computing Group at Microsoft Research Asia.1
  2. Became a research lead in Microsoft's Natural Interactive Services Division.1
  3. Became a Principal Researcher in the Natural Language Processing Group at Microsoft Research, Redmond.1
  4. Co-authored the DSSM paper on learning semantic models for web search from clickthrough data (CIKM).3
  5. Posted "Neural Approaches to Conversational AI" with Michel Galley and Lihong Li.5
  6. Co-authored the LLaVA-Med paper on a biomedical vision-language assistant.4
  7. Co-authored the Phi-3 technical report introducing phi-3-mini.2

Other documented roles

Roles the sources document that are not confirmed as current.

  • Distinguished Scientist and Vice President; co-founder and long-time leader of the Deep Learning Group, Microsoft Research1

Honors and recognition

  • IEEE Fellow · Institute of Electrical and Electronics Engineers1
  • ACM Fellow · Association for Computing Machinery1
  • ACL Fellow · Association for Computational Linguistics1

Where they work

Suggest a correction or removal

Profiles cover public professional work only, and each statement cites a source. If something is out of date, missing or wrong, tell us and link a source. People can also ask for their profile to be corrected or removed.

References

6 sources · profile reviewed Oct 21, 2026 · data updated Oct 21, 2026

  1. Jianfeng Gao at Microsoft ResearchMicrosoft Research · Official page · accessed Oct 21, 2026
  2. Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone (arXiv 2404.14219)arXiv · Paper · published Apr 22, 2024 · accessed Oct 21, 2026
  3. Learning Deep Structured Semantic Models for Web Search using Clickthrough DataMicrosoft Research · Paper · published Oct 2013 · accessed Oct 21, 2026
  4. LLaVA-Med: Training a Large Language-and-Vision Assistant for Biomedicine in One Day (arXiv 2306.00890)arXiv · Paper · published Jun 1, 2023 · accessed Oct 21, 2026
  5. Neural Approaches to Conversational AI (arXiv 1809.08267)arXiv · Paper · published Sep 21, 2018 · accessed Oct 21, 2026
  6. microsoft/Phi-3-mini-4k-instruct model card (license MIT)Microsoft (Hugging Face) · Model card · accessed Oct 21, 2026