Mahi Shafiullah 🏠🤖 (@notmahi) / X

Mahi Shafiullah 🏠🤖

963 posts

Mahi Shafiullah 🏠🤖

@notmahi

Trying to understand the emergence of generally intelligent robotic behavior at @berkeley_ai. Previously @CILVRatNYU @MIT & @Apple AI/ML fellow.

New York, NY

Joined October 2010

Pinned
Mahi Shafiullah 🏠🤖
@notmahi
Feb 12
Why buy a robot when you can build your own? Meet YOR, our new open-source bimanual mobile manipulator robot – built for researchers and hackers alike for only ~$10k. 🧵👇
00:00
38K
Mahi Shafiullah 🏠🤖
@notmahi
Nov 28, 2023
Proud to announce Dobb·E: the next step in home robot system that I was working on for the past 3 years. We have visited 10 homes, learned 100+ tasks, and we are just getting started! And we fully open-sourced it all, hardware, models, and software: dobb-e.com 🧵
00:00
164K
Mahi Shafiullah 🏠🤖
@notmahi
Sep 9, 2024
Robot Utility Models (RUMs) enable basic tasks – door opening, drawer opening, object reorientation, etc. – at ~90% accuracy without ANY finetuning (i.e. zero-shot) in unseen new environments. Fully open source!!! models, data, code & hw. We think this is super exciting, why?👇
00:00
90K
Mahi Shafiullah 🏠🤖
@notmahi
Oct 12, 2022
How can we train data-efficient robots that can respond to open-ended queries like “warm up my lunch” or “find a blue book”? Introducing CLIP-Field, a semantic neural field trained w/ NO human labels & only w/ web-data pretrained detectors, VLMs, and LLMs mahis.life/clip-fields
00:00
Mahi Shafiullah 🏠🤖
@notmahi
Sep 2, 2025
Happy to share that I'm a Dr. now! 🎓🎉 I'm indebted to @LerrelPinto for his 🔥 advice & mentorship, and an unforgettable five years. Next up, I'm starting a postdoc with @JitendraMalikCV, spending time at @berkeley_ai & @AIatMeta. Would love to say hi if you're in the bay area!
16K
Mahi Shafiullah 🏠🤖
@notmahi
Jan 23, 2024
Wouldn’t it be nice if you could bring a robot home, give it a video of your room, and immediately start asking it to move objects around? Turns out, now you can! Introducing OK-Robot, a zero-shot language-specified pick & drop system that we built with exactly ZERO training! 🧵
00:00
57K
Mahi Shafiullah 🏠🤖
@notmahi
Mar 4, 2023
Delighted to share that I've been awarded the 2023 @Apple Scholars in AI/ML PhD fellowship for my work in Machine Learning for robots: machinelearning.apple.com/updates/apple-… Huge part of the credits go to @LerrelPinto for the invaluable advising and support. Thank you!
machinelearning.apple.com
Announcing the 2023 Apple Scholars in AIML
Apple is proud to announce the 2023 recipients of the Apple Scholars in AIML PhD fellowship.
13K
Mahi Shafiullah 🏠🤖
@notmahi
Aug 1, 2024
The "Supervised Policy Learning for Real Robots" tutorial at #RSS2024 is now public! supervised-robot-learning.github.io We're trying to democratize the "folk knowledge" of training real robot policies, and cover the fundamentals of imitation, recent algorithms, & practical considerations.
10K
Mahi Shafiullah 🏠🤖
@notmahi
Nov 8, 2024
If you want robots that can just live with you & help 24/7, it needs to build & update its memory on the fly. Current semantic memory representations like VoxelMap from OK-Robot can't change with the world. That's why we built DynaMem: dynamic memory for a changing, open world!
00:00
30K
Mahi Shafiullah 🏠🤖
@notmahi
May 21, 2025
Morning, #ICRA2025 @ieee_ras_icra! Bring something small 🍋🍑 and have our Robot Utility Model pick it up at our EXPO demo today from 1-5 PM, between hall A2/A3! Talk and poster is right before, 11:15-12:15 in room 411. Also, DM if you want to chat 🤖 for the messy, real world!
00:00
14K
Mahi Shafiullah 🏠🤖
@notmahi
Jan 4, 2024
"Imitation learning just works" should probably be one of the biggest takeaways from the last year in robotics. Congrats @tonyzzhao @zipengfu and @chelseabfinn on this amazing work! Let's make the setup more compact so I can bring it to my (grad student sized) home :D
Chelsea Finn
@chelseabfinn
Jan 3, 2024
Meet Mobile ALOHA! 🤖 What’s new: - low-cost but widely-capable teleop platform 🦾🛞 - imitation learning just works Paper/Code/Videos: mobile-aloha.github.io Led by @zipengfu and @tonyzzhao
00:00
19K
Mahi Shafiullah 🏠🤖
@notmahi
Dec 16, 2023
Had a fun day of demoing Dobb•E with @HarithejaE at NeurIPS! Tomorrow we will be at the robot learning workshop (hall B2), so stop by for some real robot fun if you missed us today :)
00:00
44K
Mahi Shafiullah 🏠🤖
@notmahi
Feb 23, 2024
OK-Robot code is now open! Check out github.com/ok-robot/ok-ro… We want to make it even better by collaborating with the community, so join our discord if you have questions, need help, or want to help discord.gg/wzzZJxqKYC
Chris Paxton
@chris_j_paxton
Feb 23, 2024
There’s nothing like seeing a robot working at home for yourself. It has to move a variety of unusual objects, like the green cactus plushie in this video, and place them intelligently in a very messy world. So on that note, we're opening up OK Robot to the community! A thread:
00:00
GitHub - ok-robot/ok-robot: An open, modular framework for zero-shot, language conditioned pick-a...
From github.com
8.3K
Mahi Shafiullah 🏠🤖
@notmahi
Jan 24, 2023
A little late to the party, but happy to share that play-to-policy was accepted @iclr_conf 23 as a notable-top-5% (Oral)! Website, paper, and code: play-to-policy.github.io By @jeffacce, @ybwrm, me, and @LerrelPinto
Lerrel Pinto
@LerrelPinto
Oct 19, 2022
Almost ♾ unlabeled data is the “secret sauce” for today's ML, but how do we use uncurated datasets in robot learning? Conditional Behavior Transformer makes sense of "play" style robot demos w/ no labels and no RL to extract conditional policies! Play-to-policy.github.io 🧵
00:00
play-to-policy.github.io
From Play to Policy: Conditional Behavior Generation from Uncurated Robot Data
Conditional real-world robotic policies from uncurated human play data
12K