Justin Yu

@uynitsuj

PhDing @UCBerkeley @berkeley_ai advised by Prof. @Ken_Goldberg. Robotics, AI | Intern @xdofai

Berkeley, CA

uynitsuj.github.io

Replies Reposts Media

Justin Yu
@uynitsuj
Jul 15
WARP-RM is now open source! Training code, checkpoints, eval suite, and the metric definitions behind the paper's sim tables: github.com/uynitsuj/WARP-… Built on the ABC (abc.bot) stack's open Mujoco simulation tele-operated dataset and eval suite, so the results
Justin Yu
@uynitsuj
Jun 30
Introducing WARP-RM: A Warp-Augmented Relative Progress Reward Model for Data Curation 🧵 We gave our t-shirt folding robot more demonstrations and it got worse. Every extra demo ended in a successfully folded shirt. The data wasn't bad. It was noisy. The policy couldn't tell
00:00
2
Justin Yu
@uynitsuj
Jul 15
Real2Render2Real-pilled mobile manipulation :D Kudos to @Lingxiao234 and @imhuanyuli for the brilliant idea and execution.
lingxiao guo
@Lingxiao234
Jul 14
Reality of mobile manipulation: 1,000 demos buy you 1 task, in 1 scene. Introducing 𝐖𝐀𝐍𝐃𝐀: a synthetic data engine that turns ONE demo into thousands of trajectories across unseen scenes — and the policy works in the open world. One demo in. An open world out. 🧵
00:00
4
Justin Yu
@uynitsuj
Jun 30
Excited to be attending ICML this year ✈️🇰🇷 and presenting CaP-X on agentic robotics (co-led with @letian_fu and @HaoruXue, among others who are not on X!) DMs are open. Let’s grab a coffee and talk shop on harnesses for embodied AI, reward modeling, or the latest in hardware!
Max Fu
@letian_fu
Apr 1
Robotics: coding agents’ next frontier. So how good are they? We introduce CaP-X: an open-source framework and benchmark for coding agents, where they write code for robot perception and control, execute it on sim and real robots, observe the outcomes, and iteratively improve
00:00
1
Justin Yu
@uynitsuj
Jun 30
Big thanks @QianzhongChen ofc standing on the shoulders of giants such as your work on Stage Aware Reward Modeling arxiv.org/abs/2509.25358 😎😎
Qianzhong Chen
@QianzhongChen
Jun 30
Big big step from value model to direct reward/adv. model. People in this area know how hard it is, congrats @uynitsuj and the team!
1
Justin Yu
@uynitsuj
Jun 30
Introducing WARP-RM: A Warp-Augmented Relative Progress Reward Model for Data Curation 🧵 We gave our t-shirt folding robot more demonstrations and it got worse. Every extra demo ended in a successfully folded shirt. The data wasn't bad. It was noisy. The policy couldn't tell
00:00
7