Brent Yi

Hello! I'm a final-year CS PhD student at the Berkeley AI Research Lab at UC Berkeley, where I'm advised by Angjoo Kanazawa and Yi Ma.

I'm currently an intern at Amazon FAR. Before, I worked with Jeannette Bohg during an MS at Stanford and Pieter Abbeel during a BA at UC Berkeley. My PhD work is supported by the NSF Graduate Research Fellowship.

GitHub  /  Twitter  /  Email

A photo of the author

Research

I'm interested in practical machine learning for real-world systems. These days, I've been thinking about generative models for action1Flow Policy Gradients for Robot Control (2026),2Flow Matching Policy Gradients (2026),4Estimating Body and Hand Motion in an Ego-sensed World (2025) and spatial perception3Cameras as Relative Positional Encoding (2025),5Reconstructing People, Places, and Cameras (2025),6Canonical Factors for Hybrid Neural Fields (2023); I also enjoy building open-source research tools.

Feel free to reach out if you're interested in chatting!

Publications

Full list on Google Scholar!

[3]
Cameras as Relative Positional Encoding
NeurIPS 2025
Ruilong Li*, Brent Yi*, Junchen Liu*, Hang Gao, Yi Ma, Angjoo Kanazawa.
 → proposes PRoPE, which you can also find in Hunyuan WorldPlay and SVSM!
[4]
Estimating Body and Hand Motion in an Ego-sensed World
CVPR 2025 (Highlight)
Brent Yi, Vickie Ye, Maya Zheng, Yunqi Li, Lea Müller, Georgios Pavlakos, Yi Ma, Jitendra Malik, Angjoo Kanazawa.
[5]
Reconstructing People, Places, and Cameras
CVPR 2025 (Highlight)
Lea Müller, Hongsuk Choi, Anthony Zhang, Brent Yi, Jitendra Malik, Angjoo Kanazawa.
[6]
Canonical Factors for Hybrid Neural Fields
ICCV 2023
Brent Yi, Weijia Zeng, Sam Buchanan, Yi Ma.

Software

I like thinking about tools and abstractions:

[9]
Type-safe CLI and config generation for Python.
Brent Yi.
GitHub stars
[10]
Solve sparse, non-Euclidean, and constrained nonlinear least squares problems in JAX.
Brent Yi.
GitHub stars
 → used in projects like EgoAllo, PyRoki, and VideoMimic!
[11]
Modular toolkit for inverse kinematics, trajectory optimization, and motion retargeting.
Chung Min Kim*, Brent Yi*, Hongsuk Choi, Yi Ma, Ken Goldberg, Angjoo Kanazawa.
GitHub stars
[12]
Lightweight framework for GPU-accelerated robot learning.
Kevin Zakka, Qiayuan Liao, Brent Yi, Louis Le Lay, Koushil Sreenath, Pieter Abbeel.
GitHub stars

Misc

I previously spent time on embedded systems and hardware design. Here's a miniature portfolio.

I'm from Milpitas, CA. Outside of research, I enjoy biking around the Bay Area and taking photos!