About

Mengyuan (Millie) Wu

Researcher • Developer • Inventor

I am a first-year Ph.D. student in Computer Science at Columbia University, advised by Prof. Xuhai “Orson” Xu in the SEA Lab. I design, engineer, and evaluate AI agentic systems that interact with people and their multimodal context to support personalization, productivity, and well-being. My research spans AI agents, multimodal sensing, and human-computer interaction. I am especially interested in building efficient, robust, and adaptive systems that can operate persistently and be meaningfully deployed and adopted in real-world settings.

I also work as a part-time full-stack software engineer at Equa Health, a CMU-funded startup. In Fall 2026, I am a teaching assistant for COMS W4995: The Future of Personal Health Assistant at Columbia. In my free time, I like to tinker with creative software and hardware projects that could bring a positive impact. Additionally, I'm a huge foodie and love to explore different restaurants in NYC.

Publications

Augmenting Human Performance with an XR Agent Learning from Online Behavior and BCI Evidence.

Li, Z.*, He, X.*, Chen, H.*, Zuo, J.*, Bai, S., Yang, B., Wu, Mengyuan, Ledner, J., Cheng, Y., Yamauchi, A., Turakhia, D. G., Feiner, S., Sajda, P.

ACM UIST 2026

MindfulAgents: Personalizing Mindfulness Meditation via an Expert-Aligned Multi-Agent System.

Wu, Mengyuan, Jiang, Z.*, Fan, Y.*, Feng, R., Dharmavaram, S., Fallon, S., Polowitz, M., Islam, B., Benson, L., Tung, I., Creswell, J. D., Xu, X.

ACM CHI 2026

Best Paper Honorable Mention Award (Top 5%)

SwEYEpinch: Exploring Intuitive, Efficient Text Entry for Extended Reality via Eye and Hand Tracking.

Li, Z.*, He, X.*, Wu, Mengyuan, Tong, Z., Wei, H., Yang, B., Feiner, S., Sajda, P.

ACM CHI 2026

PAWS: Empowering Everyday Cannabis Use Disorder Support through a Personalized AI Digital Pet on Smartwatches.

Jiang, Z., Wu, Mengyuan, Zou, R., Xu, S., Macmaunus, E., Liao, S., Zhang, P., Wang, D., David, J. L., El-Bassel, N., Mamykina, L., Levin, F. R., Sultan, R., Xu, X.

ACM CHI 2026 Workshop

Best Poster Award

Manuscripts under review

"Live Your Day, Twice": A Multi-Modal Agentic Mental Rehearsal Framework for Closing the Intention-Action Gap.

Wu, Mengyuan, Guan, X.*, Nam, Y.*, Chung, R., Wang, K., Wang, Z., Xu, Z., Babbit, E., Bloom, P., Funkhouser, C., Mishra, S., Xu, X.

Submitted to ACM CHI 2027

Deco: Making Physical Companions Pervasive through a Dual-Embodiment Framework.

Jiang, Z.*, Wu, Mengyuan*, Zou, R., Xu, S., Xu, X.

Submitted to ACM CHI 2027

ExTENDS: Empowering Domain Experts to Build Personalized AI Proxies for Asynchronous Patient Care.

Chung, R., Li, M.*, Wu, Mengyuan*, Pollack, A., Pratt, W., Mamykina, L., Xu, X.

Submitted to ACM CHI 2027

TimeSRL: Generalizable Time-Series Behavioral Modeling via Semantic RL-Tuned LLMs — A Case Study in Mental Health.

Fan, Y., Wu, Mengyuan, Chen, Q., Liu, X., Yang, Y., Zhang, Z., Nepal, S., Nie, J., Jiang, X., Xu, X.

Submitted to IMWUT 2026 · Under major revision

Agents’ Last Exam.

Sun, Y., et al. (Mengyuan Wu, co-author)

Submitted to NeurIPS 2026 Evaluations and Datasets Track

Sharpen Before You Adapt: Data-Free Entry-State Sharpening for Test-Time Reinforcement Learning.

Zhang, Z.*, Selvendran, V.*, Wu, Mengyuan

Submitted to NeurIPS 2026 Workshop

Experience

Research

2024-Present

Research Assistant, SEA Lab

Columbia University · New York, NY

Mentor: Xuhai "Orson" Xu

2024

Research Assistant, LIINC and CGUI

Columbia University · New York, NY

Mentors: Steve Feiner and Paul Sajda

Industry

2025-Present

Full-Stack Software Engineer (Part-time)

Equa Health · CMU-funded startup

2021-2023

Vice President of Investment Research

Evercore Partners

Teaching

  • Fall 2026: COMS W4995 The Future of Personal Health Assistant, Teaching Assistant
  • Fall 2024: COMS W4995 Applied Machine Learning, Course Assistant
  • Spring 2019: IEOR E3609 Advanced Optimization, Course Assistant

News

  • Fall 2026: I'm a teaching assistant for The Future of Personal Health Assistant (BINF 4070 / COMS W4995) at Columbia University.
  • 2026: Our PAWS project was showcased at Columbia's 2026 AI at VP&S Showcase, with Zhihan Jiang presenting the team's demo.
  • 2026: PAWS received the Best Poster Award at the Everyday Wearable for Personalized Health and Well-Being Workshop at ACM CHI 2026!
  • 3/14/2026: Our paper was featured in SciPaperMill's article, "Mental Health: Unpacking the AI Revolution in Therapeutic Support."
  • 3/8/2026: MindfulAgents received an Honorable Mention Award at ACM CHI 2026!
  • 2/26/2026: Our U.S. provisional patent application for "Method for Generating a Meditation Protocol" was filed (No. 63/991,687), with Equa Health, Inc. as assignee.
  • 2/11/2026: Two papers were accepted to ACM CHI 2026, and I'll be attending the conference in Barcelona in April.

Projects

Expert-Aligned Personalized AI Meditation

Expert-Aligned Personalized AI Meditation

Research to Product • Patent Application 63/991,687

A multi-agent system that generates personalized meditation experiences under an expert-aligned framework.

Law of Attraction AI Journalling

Law of Attraction AI Journalling

HooHacks Intel AI Track, 2nd Place

A context-aware journalling assistant built to help users articulate and revisit future-oriented intentions.

View project
EEG-Art for Mindfulness

EEG-Art for Mindfulness

NeuReality AR/VR Competition, 1st Place

A generative art system translating EEG signals from mindfulness practice into expressive visual outputs.

View project