I'm interested in research questions that will still matter five years from now.
I build interpretable, human-centered foundation models — and the systems and robots that bring them into the real world.
(1) Interpretable-by-design, mechanistically grounded foundation models
I develop foundation models whose internal representations are structured and inspectable, and whose reasoning, planning, and dynamics can be examined and controlled.
(2) Human-AI collaboration and deployment in high-stakes healthcare settings
I build collaborative, deployed systems for precision and public health — speech and language disorders such as dyslexia and primary progressive aphasia, and population-level language screening.
(3) Human-centered speech AI systems
I build speech AI that is grounded in human cognition and physiology (perception, production, physics) and that keeps people in the loop through interactive labeling, feedback, and supervision.
(4) Inverse problems for intelligence modeling
I work to find intelligence in AI models, aiming to steer, predict, and understand it.
News & Updates
Released the Eurandel AI preview: a multimodal agent towards an empathetic, personalized assistant for life. Currently in beta for healthcare and education, non-commercial testing only. Looking for all-purpose collaborators! eurandel.ai →
I launched ClawVoyage 🦞 as an OPC in China. Within six months, it reached 10M users and ~$30M+ in transaction volume, and has since undergone a merger. I also received 50+ speaking invitations from Shenzhen’s 🦞 community. Stay tuned for what comes next!
We host the Bay Area Speech AI for Health, Education, and HCI Workshoptwice a year, connecting academia, industry, healthcare, education, and venture capital to advance human-centered speech intelligence. Join us to present, exchange ideas, or collaborate!
As approved by the California State Government, starting from 2025 public schools will adopt our language screener, where I developed the first and state-of-the-art speech dysfluency transcriber UDM / SSDM, serving 1 million kids! Read the report →
Selected Academia Publications
Human-Centered Speech AI System
HuPER: A Human-Inspired Framework for Phonetic Perception
Chenxu Guo* (co-first), Jiachen Lian*‡ (co-first), Yisi Liu, Baihe Huang, Shriyaa Narayanan, Cheol Jun Cho, and Gopala Krishna Anumanchipalli,
2026 arXiv. ‡ Project Lead. Human-centered design that elicits 100× greater data efficiency than scaling-based methods.HuPER spawned state-of-the-art phonetic modeling, adopted and scaled up in Qwen3.8-Omni-Flash to power the world's strongest language-learning AI.
Invited to give a talk at ChangeLing Lab.
[Code][🤗 36,076 downloads]
• Core contributor to Meta's open-weight LLaMA models — total Hugging Face downloads:
🤗 LLaMA 3.2216,187,015
· 🤗 LLaMA 3.312,902,451
· 🤗 LLaMA 410,386,581
Industrial
Meta AI, CA, USA
Visiting Researcher • Sep 2024 to June 2026
LLaMA Team, Seamless Team; also with Abdelrahman Mohamed. LLaMA 4: duplex pre/post-training
Introduction to Deep Learning, CMU LTI , PA, USA
Teaching Assistant (Head) • Fall 2020
Instructor: Prof. Bhiksha Raj
Education
UC Berkeley, U.S.
Ph.D. in EECS • Aug. 2021 to Present
Carnegie Mellon University, U.S.
M.S. in ECE • Sept.2019 to Dec. 2020
Zhejiang University, China
B.Eng. in EE • Aug. 2015 to June 2019
Awards
2026 EECS Evergreen Graduate Research Award 2025 Radical Ventures AI Founders Grant ($350K) 2025 ASRU AI4CSL Best Paper 2024 Meta AI Mentorship (AIM) Program — 2-Year PhD Funding 2024 NeurIPS Scholar Award 2024 Sevin Rosen Funds Award 2023 ASRU Best Paper Nomination 2021 Berkeley Golden Fellowship
