PeopleofSuperintelligence

Researcher

Guoqing Zheng

Guoqing Zheng is a language technology researcher whose work covers reasoning in small language models, synthetic training data, and decentralized mixture-of-experts systems. He studied computer science at Shanghai Jiao Tong University and earned an M.S. and Ph.D. at Carnegie Mellon. From 2018 to 2025, he served as a principal researcher at Microsoft Research. In 2023 he coauthored Orca 2; in July 2024 he coauthored AgentInstruct, which generated 25 million instruction pairs for language-model training. In 2025, he coauthored a paper introducing DDOME, a decentralized mixture-of-experts framework, and since 2025 he is a researcher at Percepta AI.

Timeline

  1. 2009–2018

    Degrees at SJTU and CMU

    He earns a B.E. in computer science at Shanghai Jiao Tong University in 2009 and an M.E. there in 2012. At Carnegie Mellon, he completes an M.S. in language technologies in 2014 and a Ph.D. in language and information technology in 2018, with Yiming Yang as his academic advisor.

  2. 2018–2025

    Principal researcher at Microsoft

    From 2018 to 2025, he serves as a principal researcher at Microsoft Research.

  3. 2023

    Orca 2 for model reasoning

    In 2023, he coauthors Orca 2: Teaching Small Language Models How to Reason. The paper teaches smaller language models to choose among strategies such as step-by-step reasoning, recall then generate, recall-reason-generate, and direct answer.

  4. 2024

    AgentInstruct training data

    In July 2024, he coauthors AgentInstruct, an agentic framework that generates synthetic data from raw text and code. The team creates a 25-million-pair post-training dataset for teaching language-model skills such as text editing, tool use, coding, and reading comprehension.

  5. 2025

    DDOME for decentralized learning

    In 2025, he coauthors “Learning to Specialize,” which introduces DDOME, a decentralized Mixture-of-Experts framework for allocating domain-specific expertise. In federated-learning tests, DDOME improves accuracy by 4% to 24% over state-of-the-art baselines.

  6. Since 2025

    Researcher at Percepta AI

    Since 2025, he is a researcher at Percepta AI. His research focuses on large-scale machine learning, natural language processing, and information retrieval.

  1. Google had the most people shaping AI every year from 2014 to 2025Our estimates of which organizations employed the most people shaping AI from 1961 to 2025, with a year-by-year chart.
  2. The United States employs an estimated 53% of people in our data with a known employment countryOur estimates of where AI careers are concentrated now and how countries’ shares of dated positions changed over time.
  3. 46% of living people in our data with known birth and employment countries work outside their birth countryOur estimate of cross-border careers in AI, from country of birth and the country of each current organization.
  4. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionOur estimates of where the people shaping AI studied and where their careers led, from education records in cited biographies.

Sources and credits

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

  1. Guoqing Zheng cs.cmu.edu. Checked 19 October 2026.
  2. Guoqing Zheng - Language Technologies Institute - School of Computer Science - Carnegie Mellon University lti-web-prod-vip.andrew.cmu.edu. Checked 19 October 2026.
  3. Guoqing Zheng — Lacuna lacuna.tiptreesystems.com. Checked 19 October 2026.
  4. Orca - Microsoft Research: Publications microsoft.com. Checked 19 October 2026.
  5. Orca 2: Teaching Small Language Models How to Reason arxiv.org. Checked 19 October 2026.
  6. AgentInstruct: Toward Generative Teaching with Agentic Flows - Microsoft Research microsoft.com. Checked 19 October 2026.
  7. Learning to Specialize: Joint Gating-Expert Training for Adaptive MoEs in Decentralized Settings proceedings.neurips.cc. Checked 19 October 2026.