Linzhan Mou
AboutWorkProjectsOnline
HI-FI

Hi, I'm Linzhan.

First-year CS Ph.D. student at Princeton, advised by Prof. Szymon Rusinkiewicz. I work on coding agents for robot learning.

More about me

A desk at night by a rainy window: a small TV showing the name Linzhan Mou, a VCR, a stack of labelled video tapes, a desk lamp and a plant.

Static version

Personal websites and portfolios are some of the last places where the Internet is still creative and fun, so I worked very hard on an interactive experience. Your computer and browser identified themselves as a device where the experience might be bad, so you've been routed to this static version.

Click here to try the interactive version anyway.

Work

Princeton University2025–now

Ph.D. in Computer Science, advised by Prof. Szymon Rusinkiewicz, in the Princeton ImageX Lab (PIXL). I work on coding agents for robot learning. pixl.cs.princeton.edu

Meta Research2026

Research Scientist Intern: video-action pretraining on egocentric videos for generalizable robot policy. ai.meta.com/research

GRASP Laboratory, Penn2024–2025

Research Intern with Prof. Kostas Daniilidis and Prof. Lingjie Liu: diverse 3D motion generation by distilling video diffusion priors (ICCV 2025 Highlight). grasp.upenn.edu

MARS Lab, Tsinghua2023–2025

Undergraduate Research Assistant with Prof. Hang Zhao: real-to-sim-to-real reinforcement learning for perceptive robot locomotion. group.iiis.tsinghua.edu.cn/~marslab

Zhejiang University–2025

B.Eng. in Robotics, Chu Kochen Honors College. GPA 3.99/4.0, ranked 1 of 121. zju.edu.cn

Publications

UniMate2026

One unified model to animate diverse skeletons from a rigged 3D asset and a text prompt. SIGGRAPH Asia 2026. linzhanmou.com/unimate

TTT-Parkour2026

Rapid test-time training for perceptive humanoid parkour: a policy fine-tuned in minutes on a reconstruction of each new course. CoRL 2026. ttt-parkour.github.io

DIMO2025

Diverse 3D motion generation for arbitrary objects from a single image. ICCV 2025 Highlight. linzhanmou.com/dimo

VR-Robo2025

A real-to-sim-to-real framework for visual robot navigation and locomotion. RA-L 2025. vr-robo.linzhanmou.com

Let Occ Flow2024

Self-supervised 3D occupancy and occupancy flow prediction from cameras alone. CoRL 2024. github.com/eliliu2233/occ-flow

Instruct 4D-to-4D2024

Editing 4D scenes as pseudo-3D scenes with text instructions, using 2D diffusion. CVPR 2024. immortalco.github.io/Instruct-4D-to-4D

Projects

  • LEGO World, LEGO CAD
  • Motion Lab, Animation
  • VR-Robo, Robots
  • OpenTracker, Analytics
  • OpenServer, Setup

Projects

  1. LEGO WorldLEGO sets as interactive CAD
  2. Motion Lab3D character animation, live
  3. VR-RoboRobot locomotion over splat scenes
  4. OpenTrackerSelf-hosted web analytics
  5. OpenServerOne-command machine setup

Online

Emaillinzhan@princeton.edu
X@LinzhanMou
GitHubFriedrich-M
Google ScholarLinzhan Mou
HomeAboutWorkProjectsOnline
HI-FI