Philipp Wu

🎉Excited to share a fun little hardware project we’ve been working on. GELLO is an intuitive and low cost teleoperation device for robot arms that costs less than $300. We've seen the importance of data quality in imitation learning. Our goal is to make this more accessible 1/n

have literally wanted this capabiltiy for mujoco for so long

Autogaze for video understanding! Up to 4-100x token savings, 19x speedup, and enables scaling to 4K-res 1K-frame videos.

With @ki_ki_ki1's help, mjlab now provides terrain normal estimation (green arrow) and foot height sensors (magenta dots). Should significantly improve rough terrain locomotion.

Applied to Claude Code and Codex OSS programs for my MuJoCo work (mjlab + related tools), but didn’t get in 😢. If anyone at OpenAI or Anthropic is open to taking another look, would love to share more about what I’m building and its impact on the ecosystem.

Amazing work by @mitchaiet! Full mocap-to-real pipeline for the G1, all open-source and powered by mjlab + MuJoCo-Warp. Amazing to see the community building on this stack 🚀🔥

:o sim2real for manipulation

Happy Friday!! mjlab v1.2.0 is out. This is our biggest release yet with 60+ PRs from 12 contributors. pip install mjlab Some highlights include: - New more powerful domain randomization module - Revamped ergonomic viewers - Cloud training via @SkyPilot - Complete doc rewrite

@snasiriany Congrats on the release @snasiriany! Very needed in the robotics community!

Proud to share the final project of my PhD, RoboCasa365! There has been so much progress in robot learning in the last couple of years but it’s starting to feel like running large-scale experimentation is increasingly out of reach for independent researchers and there is no consensus yet on benchmarking in the field. During my PhD I wanted to build something that would allow myself and other researchers to study robot learning on large datasets in a meaningful way. So I dedicated my time to building RoboCasa, a large-scale simulation framework for training and benchmarking generalist robot policies. We released the original framework in 2024 and today we are releasing a major new release, RoboCasa365. Compared to the original release, RoboCasa365 feels a lot more like a “full stack” simulation framework: - 2500 kitchen scenes - 365 everyday tasks - 600+ hours of teleportation data and 1600+ hours of synthetically generated trajectories - Benchmarking on sota VLA models By current industry standards, 600 hours of teleportation data is considered modest, but I think this is a good sweet spot of data to study how well robot foundation models can adapt to downstream applications. Right now the benchmark is far from solved. This makes it a useful tested to develop the next algorithms and architectures to push the boundaries on robot learning, be it VLAs, world models, RL algorithms, etc. There is a lot of work left to push generalization, reliability, and throughput for general purpose robots. I was incredibly lucky to have the support of my adviser @yukez, who gave me all the creative freedom, resources, and time to build RoboCasa. I am also very fortunate to work with two hardworking and passionate students, @abhirammaddukur and my brother @SepNasiriany. We spent countless long nights together on this project and it was really fun working as a lean team. I also want to thank @bgxc and the team at @LightwheelAI for being a major supporting force in sourcing assets and collecting the data that we used in this project. Also thank you to the RPL lab and Nvidia for supporting our work. You can now check out RoboCasa365 at robocasa.ai!

@brenthyi who worked on FPO/FPO++ is finishing his PhD and going on the job market 😭✨ He is also the person behind viser, pyroki, egoallo, jaxls, tyro and more! I can't express how amazing it is to have Brent on your team..! Any team would be incredibly lucky to have him!!

:o

🚀 DreamZero training code is LIVE — train your own WAM (aka VAM)! 🔧 Replicate DROID from-scratch training 📊 Run evals on sim (DROID-Sim, MolmoSpaces, Polaris) & real-world (RoboArena) No 2 GB200s for real-time inference? No problem — let NVIDIA carry that burden 💪. Sign up for our API and jump into prompting new tasks! (e.g. "fan the burger" 🍔, totally unseen verb/task from DROID) Coming soon: new embodiment/robot fine-tuning initialized from our DreamZero-AGIBot checkpoint. Stay tuned! 🤖 🔗 github.com/dreamzero0/dre…

Continuing on the journey to accelerate towards an optical future. Come help us connect and move things with optical photons 🥰 working with an amazing team!

Introducing Mesh

Really excited to see the creativity and community engagement around mjlab 😍 Here’s a thread of what people are building 🧵👇

So excited to try out the sysid tool box! Hint: sysid is the actual secret sauce for sim2real

@qiayuanliao incredible!

One of my favorite robot clips (filmed Oct 2025). You can train any crazy full-body motions like this with our open-source stack without changing any parameters. whole_body_tracking: github.com/HybridRobotics… mjlab: github.com/mujocolab/mjla…

New project! Flow Policy Gradients for Robot Control tldr; a simple online RL recipe for training and fine-tuning flow policies for robots co-led w/ @redstone_hong: hongsukchoi.github.io/fpo-control

Introducing DreamZero 🤖🌎 from @nvidia > A 14B “World Action Model” that achieves zero-shot generalization to unseen tasks & few-shot adaptation to new robots > The key? Jointly predicting video & actions in the same diffusion forward pass Project Page: dreamzero0.github.io 🧵 (1/10)

Definitely a fun project. My biggest takeaway is that varying the camera viewpoint really makes the learning more robust.