# Andi Peng

> Andi Peng is an AI researcher and co-founder of humans& whose path through public policy, Microsoft Research, MIT robotics, and Anthropic has been shaped by a recurring question: how should intelligent systems learn what people mean and value? She left humans& in July 2026 after helping launch its human-centered AI agenda.

- **Role:** AI researcher and founder
- **Organizations:** humans&, Anthropic, MIT Computer Science and Artificial Intelligence Laboratory, Microsoft Research, Yale Jackson School of Global Affairs, Defense Innovation Unit, White House Office of Science and Technology Policy, National Institute of Standards and Technology, Yale University
- **Education:** B.S. in Cognitive Science; B.A. in Global Affairs, Yale University, M.S. in Computer Science, Massachusetts Institute of Technology, Ph.D. studies in Computer Science (on leave as of 2024), Massachusetts Institute of Technology
- **Known for:** 2017 Truman Scholar., Lead author of the 2023 ICML paper Diagnosis, Feedback, Adaptation on interactive robot policy adaptation., Co-author of publications on human-AI teamwork, robot representation alignment, and language-guided abstraction.

## Career timeline

- **2015** — Head teaching assistant for Yale’s introductory computer science course.
- **2017** — Named a Truman Scholar while studying cognitive science and global affairs at Yale.
- **2018** — Worked on federal science policy in Washington and joined Microsoft Research’s AI Residency.
- **2020** — Began doctoral study at MIT CSAIL, focusing on human feedback and AI systems.
- **2023** — Co-authored work on helping users diagnose and correct robot failures; earned an MIT master’s degree.
- **2024** — Joined Anthropic, later working on reinforcement learning and Claude post-training.
- **2025** — Co-founded humans& with Eric Zelikman, Yuchen He, Georges Harik, and Noah Goodman.
- **2026-01** — Humans& announced a $480 million seed round.
- **2026-07** — Publicly announced her departure from humans&.

## Achievements

- 2017 Truman Scholar.
- Lead author of the 2023 ICML paper Diagnosis, Feedback, Adaptation on interactive robot policy adaptation.
- Co-author of publications on human-AI teamwork, robot representation alignment, and language-guided abstraction.
- Worked on reinforcement learning and post-training of Claude 3.5 through 4.5 at Anthropic, according to her personal site.
- Co-founded humans&, which raised $480 million in seed funding in January 2026.

## Latest updates

- **2026-07** — Announced on X that she was leaving humans& and would cheer on the team.
- **2026-01** — Spoke with TechCrunch about co-developing the humans& model and product as the company announced its $480 million seed round.

## Links

- Website: https://andipeng.com/
- LinkedIn: https://www.linkedin.com/in/andipeng
- Twitter/X: https://x.com/TheAndiPenguin

---

Profile page: https://yespress.io/andi-peng
Published by YesPress — https://yespress.io
Last updated: 2026-09-29