
plan 24, exec 8. robot learning @ gt 🤖 | adv. @danfei_xu
gatech
Joined June 2020
- Tweets49
- Following223
- Followers197
- Likes216
Pinned Tweet
🤖 Scaling deployment in robotics can be difficult and expensive. How can we properly serve a VLA/WAM to 10+ robots concurrently from a single GPU?
Introducing Armory, a system for serving large robot policies to a fleet of robots from the cloud/edge. Through real-time scheduling of action chunks, Armory boosts task throughput over baselines by up to 18% on heterogeneous fleets in the real world.
🧵1/6
Rohan Bansal retweeted
Check out our work on representing “motion” as 3D scene flow with a video generation backbone. XYZ->RGB trick for using pretrained WAN VAE might be my favorite idea 😆
Learned so so much on this project, great collaborators and mentors: @jerthesquare_ @NamraByTheSea @agarwal_ayush05 @animesh_garg
Rohan Bansal retweeted
Most world action models learn to act by predicting videos.
We find that predicting how the scene moves in 3D performs better.
Introducing FloMo: a world action model that uses 3D motion as the bridge between video pre-training and robot control.
national ground stop on friday the 13th, i suppose it’s fitting
anyway +1 day for corl coffee chats
2
10
1,095
super cool application of 3D flow from Shuo and co.!
What should transfer from a human video to a robot?
Human2Any learns how objects interact, then adapts that knowledge through test-time reasoning and planning.
Accepted to #CoRL2026! I'll be presenting in person next week—come say hi!
human2any.github.io (1/n)
I'll be presenting Armory at CoRL next week! I can't wait to see what Austin has to offer and make lots of new friends. Send me a DM if you'd like to chat about robot inference, or co-designing representations, policies, and control for robot learning!
🤖 Scaling deployment in robotics can be difficult and expensive. How can we properly serve a VLA/WAM to 10+ robots concurrently from a single GPU?
Introducing Armory, a system for serving large robot policies to a fleet of robots from the cloud/edge. Through real-time scheduling of action chunks, Armory boosts task throughput over baselines by up to 18% on heterogeneous fleets in the real world.
🧵1/6
2
12
430
🐐
I'm excited to share that I've joined @physical_int as a Member of Technical Staff Intern.
General-purpose robots will be a defining problem of the next decade, and I can't wait to work on it with such an incredible team.
9
940
Rohan Bansal retweeted
Today, Dream Labs is coming out of stealth with Dream-1, a robot foundation model that controls bimanual robots it has never seen, on tasks it was never shown, in places it has never been, from $3/hour.
Dream API is live for everyone, today. 🧵
Rohan Bansal retweeted
I know teleoperation is old these days but it feels magical when you could teleoperate a humanoid just by wearing a pair of classes (bonus: the robot also sees from glasses too!)
Of course we don't stop here --- we are working on a system to train robots with everyday human experience, without any teleop data. More to come!
Kudos to the team Chuye Zhang @DanielZhenyang @ChuizhengKong
Yay!!! Congrats @kexinrong 🎉
Excited to receive the 2nd Annual Google ML and Systems Junior Faculty Award, recognizing our work in ML systems and data applications!
Thanks to Google for supporting our research. Looking forward to sharing more from our lab soon!
4
673
Rohan Bansal retweeted
Students sometimes ask me if it still makes sense, in this accelerating age, to pursue a PhD in AI.
Perhaps counterintuitively, I think it's a great time to do so.
I wrote up some thoughts on this here: web.mit.edu/phillipi/www/wri…
Lots of robot learning problems can be solved much faster with a smartly chosen and well tuned controller, or with good async inference infra.
Super cool, will be attending!
Robot-learning folks: have you ever found that a seemingly minor system choice beneath the policy—controller tuning, hardware selection, or inference infrastructure—made a much bigger difference to your robot’s performance than an architecture or algorithm tweak? How do you usually isolate and report that effect?
That’s the question behind our Workshop on Everything Beneath the Policy at #CoRL2026.
7
432
Addressing the individual components != addressing the systems problem. They really highlight the importance of good model-hardware-controller co-design here, through these very impressive demos.
This will be the hardest challenge for frontier LLM wrappers to overcome!
happy to share that Armory will be presented at #CoRL2026. See you all in Austin in November! Both @davidhe137 and I will be there, reach out if you would like to chat!🦾
btw, cool walkthrough video we took of a “lego sorting factory” in action: 14 robots scheduled onto 1 cloud GPU at 30hz with train-time RTC!
🤖 Scaling deployment in robotics can be difficult and expensive. How can we properly serve a VLA/WAM to 10+ robots concurrently from a single GPU?
Introducing Armory, a system for serving large robot policies to a fleet of robots from the cloud/edge. Through real-time scheduling of action chunks, Armory boosts task throughput over baselines by up to 18% on heterogeneous fleets in the real world.
🧵1/6

















