CS PhD student at @USCViterbi + intern @nvidia | prev intern @Cornell @SandiaLabs | undergrad @UTCompSci
Austin, TX
Joined August 2013
- Tweets274
- Following692
- Followers786
- Likes1.2K
Pinned Tweet
Robots are deployed for long periods of time, but how can they answer questions and generate goals based on their long-horizon history?
During my internship at #NVIDIA, we built ReMEmbR, a retrieval-augmented memory for embodied robots. 1/8 🧵
nvidia-ai-iot.github.io/reme…
A super cool way to scale data generation in simulation!
Super cool video too! Great work @neel_jawale
Multi-arm robots can do more together, but teaching them to coordinate carefully is hard.
Introducing ScheduleGen, an automated data gen system for multi-arm manipulation.
We evaluate it in 28 sim and 8 real-world tasks including real2sim2real.
njawale42.github.io/Schedule…
Abrar Anwar retweeted
We’re excited to announce the official release of NVIDIA Isaac Lab 3.0. 🎉
Isaac Lab 3.0 brings a new modular and lightweight architecture along with major new capabilities across the robot learning workflow: multi-physics and multi-rendering support, training at scale, and streamlined deployment on real hardware.
With Newton integration, the new OVRTX renderer, RL post-training, and LEAPP deployment, there’s a lot to explore.
Here’s a look at what’s new 🧵
(1/7)
Abrar Anwar retweeted
Automated robot demonstrations are often difficult to imitate because plans produce widely varying motions and visual observations. DemoStream is a new algorithm that optimizes demonstrations for learnability without task-specific human demonstrations.
Abrar Anwar retweeted
(1/6) From @GoogleDeepmind’s AlphaGo to AlphaZero, we saw machines find moves that reached beyond what we’d imagined for them. No one taught them those strategies. They emerged from reinforcement learning: playing, failing, and trying again under a single binary task reward for success or failure. Inspired by them, we couldn’t help but wonder: what might a robot discover if we gave it that same freedom to learn?
That question led us to AlphaNIST. Starting tabula rasa, with that same binary reward plus only a simple control cost, a single policy solves all 20 @NIST assembly tasks at a 98.2% success rate in simulation. More surprising still, AlphaNIST can complete 10 different tasks in a row, receiving that reward only once, after all 10 are done.
Just as AlphaGo’s strategies surprised its creators, AlphaNIST discovered dexterous strategies we couldn’t have come up with ourselves.
Paper: arxiv.org/abs/2610.13354
Website: alpha-nist.github.io
Abrar Anwar retweeted
(1/5) Everyone says dexterous manipulation needs real-world data. These policies were trained with none. Our robots mesh gears, thread nuts, and insert pegs zero-shot from RGB, and the same recipe works for locomotion across terrains much harder than before.
The missing piece wasn't a better RL algorithm, but a small tweak to what the robot practices.
Today we're releasing Success-Guided Sampling (SGS), a simple change to the task sampler that unlocks new levels of dexterity and agility with large-scale, end-to-end sim2real RL.
No demonstrations, no per-task reward tuning. Just PPO at scale.
Check out the policies yourself at sgs-rl.github.io! All videos at 1x speed! 🧵 🔈
Abrar Anwar retweeted
Runtime monitoring is key for robot deployment. Existing methods rely on task-specific failure models or policy-dependent uncertainty, requiring retraining or recalibration across tasks and policy updates.
In Reward-DAgger, we ask: can we build a gate that transfers across both?🧵
Abrar Anwar retweeted
We’re on a mission to build a billion AI models, not trained by us, but by you.
Introducing @bryel_labs we help your company build its in-house AI lab and give you the software to make model development and continual improvement simpler. We are backed by @ycombinator and have trained models at Apple, Framer, and F500s.
We’ve enabled companies to train:
• Classification models that are 20x faster
• Search models that cost 5x less.
• Long horizon agent models that match the frontier at 5-8x lower cost.
Are you interested in partnering with us? Reach out.
bryel.ai/
Abrar Anwar retweeted
You can't dump terabytes of robotics data onto a laptop and point an agent at it, but that’s what we do with most smaller shapes of data
We made working with physical data as easy as working with CSV files.
Today, we are launching Interlatent Dataroom
Abrar Anwar retweeted
Introducing GameBoyWorlds, a new challenge testbed for self-improvement in embodied video games:
Paper: arxiv.org/abs/2609.32093
Website: dhananjayashok.github.io/Gam…
Abrar Anwar retweeted
Human–Robot Dialogue @ IROS 2026: one week to go! 🤖💬
In just one week, we're bringing together researchers from robot learning, human–robot interaction (HRI), natural language processing (NLP), computer vision, and cognitive science to identify and articulate the open problems in human–robot dialogue.
🗓 Thursday, October 1 · 8:30 AM – 12:30 PM
📍 Room 412, Level 4
Invited talks:
• @imankitgoyal (NVIDIA/UM) - From Ambiguous Instructions to Grounded Actions: Reasoning for Robot Foundation Models
• @jacobandreas (MIT) - Just Asking Questions
• @_jessethomason_ (Georgia Tech) - Grounding Language in Actions, Multimodal Observations, and Robots
• @BaharIrfan_ (Familiar Machines & Magic) - Detecting User Enjoyment in Conversations to Recover from Failures
Panel: The future of human–robot dialogue with Jesse Thomason, Bahar Irfan, @GabrielSkantze (KTH · Furhat Robotics), and @matt_marge (DARPA).
Full program: human-robot-dialogue.github.…
Abrar Anwar retweeted
We’re hiring 2027 Applied Research Interns in robot learning at #NVIDIA! Join me and the Isaac Loco-Manipulation team to define and scale the next generation of humanoid robot learning!
Work on GR00T and Cosmos foundation models, hybrid simulation, real-to-sim-to-real, dexterous manipulation, and whole-body control—bringing research ideas to real world impact.
👉 Apply: docs.google.com/forms/d/e/1F…
#NVIDIA #Robotics #Internship
Abrar Anwar retweeted
#NVIDIA Applied Research Internships for 2027 🤖
Bring your research to life on real humanoid robots.
I’m hiring interns to join NVIDIA’s Isaac Loco-Manipulation team. You’ll own an applied research project at the intersection of humanoid robot learning, dexterous manipulation, whole-body control, real-to-sim-to-real, and GR00T and Cosmos foundation models—taking ideas from early experiments to real robot capabilities.
We’re looking for PhD / master students with equivalent experience and strong robotics research or engineering backgrounds. Publications at CoRL, RSS, ICRA, CVPR, or related venues are a plus.
Excited to build robots that learn and act in the physical world? Please submit your information through this application form docs.google.com/forms/d/e/1F…
Please share with anyone who might be a great fit!
#Nvidia #Robotics
We're hosting a super fun challenge at our SPIN Workshop at CoRL: the Child-Robot Dexterity Challenge. More below!
🚨 Call for participation: the Child–Robot Dexterity Challenge at the CoRL 2026 𝗦𝗣𝗜𝗡 𝗪𝗼𝗿𝗸𝘀𝗵𝗼𝗽! Bring your robot, or run on ours. Apply by Oct 5.
🧒🤖 Robots and a child will take on the same everyday tasks, side by side, live on stage. The demonstration offers a simple way to see what today’s robots can already do and where there’s still room to improve.
We make it easy to join:
- Bimanual YAM setup (MolmoAct 2 style) + a dexterous Sharpa hand setup, provided on stage.
- Bring your own hardware (subject to approval).
🌐 Website: spin-workshop.github.io/#con…
🗣️ Speakers & panel from Stanford, Columbia, TU Darmstadt, and AMI Labs.
📍 Nov 12 · Austin, TX
#CoRL2026
My team at @NVIDIARobotics is hiring PhD research interns in our Santa Clara (US) and Shanghai (China) offices! We work broadly on robot learning from RL post-training, WAMs, real2sim2real, dexterity, and more!
Fill out this form if interested!
forms.gle/ZtbNJRPCw94Nkdf99
Abrar Anwar retweeted
🎥🤖 Help turn human video into the next breakthrough in robot learning.
Join the Video to Data (V2D) Challenge across three tracks: 4D reconstruction, robotic grounding, and egocentric video-to-policy transfer.
Participate to:
🚀 Build on provided datasets, starter tools, and an end-to-end baseline.
🏆 Compete for track awards and Industry Innovation Awards.
🎤 Earn an invitation to present at CoRL 2026 as a top team.
📝 Contribute as a co-author to the post-challenge technical report with a valid final submission.
Top solutions will also receive large-scale evaluation, testing how well your approach generalizes beyond the challenge benchmark.
📅 Launches September 21. Registration is open!
👉 Explore the challenge and register nvidia-isaac.github.io/video…
We’re excited to welcome our judging panel:
🌟 @YunzhuLiYZ — Columbia University, SceniX & World Labs
🌟 Louis Lian — Lightwheel
🌟 @ChangliuL — Carnegie Mellon University
🌟 @haosu_twitr — Fudan University
#Robotics #EmbodiedAI #NVIDIA
My team at NVIDIA is hosting the Video to Data Challenge this year at CoRL. Check it out!
What could robots learn from the videos we already have?
Join the Video to Data (V2D) Challenge to help turn human demonstrations into robot skills for 3 tracks:
🔹 4D Reconstruction
🔹 Robotic Grounding
🔹 Egocentric Video-to-Policy.
You'll have the opportunity to:
🏆 Earn recognition at #CoRL2026, where the winning solutions will be announced
🛠️ Receive evaluation at scale, with top teams invited to integrate their code into a maintained repository
📋 Contribute to the challenge’s technical report
Register here by September 21, 2026:
nvda.ws/4yGsIcP
Abrar Anwar retweeted
Robot actions can be represented as motion in pixel space.
NVIDIA Research introduces Hydra-0, a generalist world model conditioned on action flow: image-plane trajectories that enable one model to learn across human hands, handheld grippers, single-arm robots and bimanual systems.
Explore the project 📄 nvda.ws/4xaO9C0
Abrar Anwar retweeted
We released DROID+, a new and improved DROID control stack from SRL. We believe robots (and hardware) should be easy to setup and easy to use, and so we updated the DROID stack to be the same. @HugoHadfield1
github.com/nvlabs/droidplus
Here's pi0.5 doing our "hello world" task:
Abrar Anwar retweeted
🚀 NVIDIA Isaac Video → Data is now open source—enabling a Real → Sim → Real workflow for robot learning.
The pipeline transforms real-world human demonstrations into simulation-ready assets and training data, then brings learned policies back to physical robots.
🎥 Segment long videos and retrieve relevant action clips with the Video Ingestion Agent
🧩 Reconstruct hand/body-object motion, depth, meshes, and 6-DoF trajectories
🤖 Retarget human motion to robot embodiments and train RL policies in Isaac Lab
🦾 Scale dexterous manipulation with using Contact
Wrench Guidance From Human Demonstration (CHORD)
📚 Access a 1,215-task motion library and 121,500 robot episodes
Real demonstrations → simulation → learned policies → real robots.
Code: lnkd.in/g3sRQ4Mm
Documentation: lnkd.in/g7RbdD4s
Dataset: lnkd.in/gnbKhYTC
CHORD: lnkd.in/gGwW3gYW
A huge thank-you to the amazing contributors whose work made this release possible.
We welcome your feedback and open-source contributions. Join us in building the future of robot learning!
#NVIDIA #Isaac #PhysicalAI #Robotics #RobotLearning #OpenSource
Abrar Anwar retweeted
What if robot hands could acquire dexterity once, then reuse it to learn downstream tasks much faster?
Introducing ADEPT: a pre-training & post-training paradigm using RL entirely in sim, then zero-shot deploying visuo-tactile policies in the real world.
🔗adept-dexterity.github.io with Dex team @NVIDIAAI