PeopleofSuperintelligence

Researcher

Hamed Khanpour

Hamed Khanpour is a researcher whose work spans natural language processing, health-related language research, and compact language models. He coauthored research on dialogue acts and online health communities before working on Orca 2, AgentInstruct, and Orca-Math. Orca 2 explored task-specific reasoning; AgentInstruct generated synthetic training data; Orca-Math applied iterative preference learning to math problems. As of October 2026, he is a Principal Program Manager at Microsoft Research and works with its ORCA team.

Timeline

  1. 2016

    Dialogue act classification

    He coauthors a 2016 COLING paper with Nishitha Guntakandla and Rodney Nielsen. It uses a deep LSTM network to classify dialogue acts in open-domain conversations.

  2. 2018

    Language work for health communities

    At the University of North Texas, he coauthors a 2018 study that uses CNN and LSTM features to identify emotional support messages in online health communities. He also coauthors a 2018 paper that combines deep neural networks with lexicon features to detect fine-grained emotions in health-related posts.

  3. 2023

    Orca 2 teaches reasoning

    In 2023, he coauthors Orca 2, which teaches small language models different reasoning techniques and aims to help them select a strategy for each task.

  4. 2024

    AgentInstruct creates training data

    In 2024, he coauthors AgentInstruct, an agentic framework that generates prompts and responses from raw documents and code. The framework creates a 25-million-pair post-training dataset for skills such as tool use and coding.

  5. 2024

    Orca-Math for grade-school math

    In 2024, he coauthors Orca-Math, a 7-billion-parameter small language model that achieves 86.81% pass@1 on GSM8K through iterative preference learning.

  6. As of October 2026

    Principal Program Manager

    As of October 2026, he is a Principal Program Manager at Microsoft Research and works with the ORCA team on compact language models and reasoning.

  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 Hamed Khanpour yet. Send us a free-licensed photo.

  1. Dialogue Act Classification in Domain-Independent Conversations Using a Deep Recurrent Neural Network aclanthology.org. Checked 19 October 2026.
  2. Identifying Emotional Support in Online Health Communities ojs.aaai.org. Checked 19 October 2026.
  3. Fine-Grained Emotion Detection in Health-Related Online Posts aclanthology.org. Checked 19 October 2026.
  4. Orca 2: Teaching Small Language Models How to Reason arxiv.org. Checked 19 October 2026.
  5. AgentInstruct: Toward Generative Teaching with Agentic Flows arxiv.org. Checked 19 October 2026.
  6. Orca-Math: Unlocking the potential of SLMs in Grade School Math arxiv.org. Checked 19 October 2026.
  7. Hamed Khanpour microsoft.com. Checked 19 October 2026.