Varun Agrawal

1.9K posts

Varun Agrawal banner
Varun Agrawal

Varun Agrawal

@varagrawal

PhD @GeorgiaTech | Robotics | CS https://t.co/QzLezFDRkB

Atlanta, GA Katılım Nisan 2009
468 Takip Edilen964 Takipçiler
Sabitlenmiş Tweet
Varun Agrawal
Varun Agrawal@varagrawal·
New paper alert! I spent 2021 building a framework for proprioceptive state estimation for all types of legged robots, and we report on some of our success. Work done with @fdellaert and @TheIHMC. Arxiv: arxiv.org/abs/2209.05644
Varun Agrawal tweet media
English
3
16
101
0
Varun Agrawal retweetledi
Georgia Tech Computing
Georgia Tech Computing@gtcomputing·
Hidden deep within @GeorgiaTech is a lab filled with some of the most advanced experimental computers in the world. This Rogues Gallery is a collection of early-stage novel computing systems that are helping shape the future. Read more: cc.gatech.edu/external-news/…
Georgia Tech Computing tweet media
English
0
3
7
577
Varun Agrawal
Varun Agrawal@varagrawal·
@MichaelAPosa Would ideas for contact estimation (as part of state estimation) be of value here?
English
0
0
0
32
Michael Posa
Michael Posa@MichaelAPosa·
Advances in contact-rich control have snuck under the radar in the last few years, but there's been major progress. Breaking from the same old workshop format, we have NO invited *talks* from the usual suspects, but instead are asking you to share your new ideas and hot takes.
John Zhang@johnzhangx

super excited to announce our full-day workshop on Contact-Rich Control and Representation (CR2) at #ICRA2026, held on June 1st in Vienna, Austria! 🕸️ website: for all information, please check out: cr2-icra2026.github.io

English
1
1
11
1K
Varun Agrawal
Varun Agrawal@varagrawal·
@johnzhangx The website is giving me a 404 error. Can you please verify GitHub pages is serving it?
English
1
0
0
52
John Zhang
John Zhang@johnzhangx·
super excited to announce our full-day workshop on Contact-Rich Control and Representation (CR2) at #ICRA2026, held on June 1st in Vienna, Austria! 🕸️ website: for all information, please check out: cr2-icra2026.github.io
John Zhang tweet media
English
2
6
13
1.7K
Robotics Papers
Robotics Papers@OWW·
Variable Elimination in Hybrid Factor Graphs for Discrete-Continuous Inference & Estimation Varun Agrawal, Frank Dellaert arxiv.org/abs/2601.00545 [cs·RO]
Robotics Papers tweet media
English
1
0
2
158
Varun Agrawal
Varun Agrawal@varagrawal·
@fdellaert It's important to note that the demo was primarily built by an MS student intern. There's a large division of labor amongst the 3 co- first authors, which is why this demo is possible.
English
0
0
0
72
Frank Dellaert
Frank Dellaert@fdellaert·
I can confirm it’s playable, and it’s a whole new way to experience results from a paper. Pretty awesome! (Using Mujoco WASM package).
Zhen Wu@zhenkirito123

Real Unitree G1 results:  • ≈3 m/s vault  • 1.25 m wall climb (~96% height)  • 48–60 s continuous multi-obstacle traversals  • Online adaptation to obstacle displacement Now playable in your browser (even on mobile): php-parkour.github.io/demo.html Powered by MuJoCo-WASM. 5/6

English
3
6
49
7.2K
Varun Agrawal
Varun Agrawal@varagrawal·
@keenanisalive The question in the last paragraph was exactly my thought process when I eschewed the corporate ladder for a career in research. This is sage advice.
English
0
0
1
70
Keenan Crane
Keenan Crane@keenanisalive·
Or to quote Robert Penn Warren: “Careerism is death.”
David Bessis@davidbessis

My recommendation for young academics is to use their intellectual sincerity (rather than institutional feedback) as a long-term navigation guide. It's OK to be a bit tactical/cynical when you're young, because the institution is putting impossible demands on you, but over the long term insincerity will destroy your creativity and your true chances of success.

English
5
8
63
9.3K
Varun Agrawal
Varun Agrawal@varagrawal·
@Slowbvrnerr_ I'm really sorry to hear that. Bureaucracy and an advisor switch led to me being delayed significantly. I had to restart with a new advisor in a new field yet I finished strong. You can also finish. Please feel free to message me.
English
0
0
0
79
Varun Agrawal
Varun Agrawal@varagrawal·
@rasmalai Ruby, zig, rust, erlang, Crystal, Rocket, Go. I can keep going...
English
0
0
0
61
Varun Agrawal
Varun Agrawal@varagrawal·
@Compass_RP @Jeb_R @ThePhDPlace I'm sorry to say but this view is a bit ill-informed. With a job you're making enough to have significant savings and handle the instability. In grad school, one trip to the ER can potentially wipe out your meager savings especially in the US where co-pay is a real problem.
English
1
0
2
31
RPCompass
RPCompass@Compass_RP·
@Jeb_R @ThePhDPlace How many students need to fight for funding? Maybe its discipline specific. And arguably, it's the most stability a person is going to have compared to any subsequent job they will hold in their careers til retirement.
English
2
0
2
625
Varun Agrawal
Varun Agrawal@varagrawal·
Despite that I am happy that I have a strong record of publication with 5 first author papers from my PhD. The results still need some more work but those are mostly due to engineering improvements.
English
1
0
1
74
Varun Agrawal
Varun Agrawal@varagrawal·
I just ran `az webapp connection create postgres-flexible ...` from the CLI, and it gave me the message: ``` unrecognized arguments: --active-directory-auth Enabled ``` Even the official CLI app is riddled with bugs, and the web SSH portal times out for me every 10 seconds.
English
0
0
0
80
Varun Agrawal
Varun Agrawal@varagrawal·
Trying to setup a basic webapp on @Azure and OMG, so many hidden steps and gotchas which the documentation doesn't address. Every time I think I've figured it out, I need to go back and enable some other service or extension. This is the worst developer experience!
English
1
0
0
94
Varun Agrawal
Varun Agrawal@varagrawal·
@amarchenkova If you're doing incremental research, then this style of writing is fine, but if you're trying to argue for a new research direction, this tends to blow up in your face no matter how fun it seems to us readers.
English
0
0
1
31
Varun Agrawal
Varun Agrawal@varagrawal·
@amarchenkova On top of that, when deep learning was black magic back in the 2010s, a lot of people advocated for a "what didn't work section" to help provide clarity. There was also suggestions of a Negative CVPR: CVPR only for negative results aka things that definitely don't work.
English
1
0
0
71
Varun Agrawal
Varun Agrawal@varagrawal·
@thegautamkamath Concur. I don't have a mountain of papers like some people but a paper I published last year had the review comment "I am confident that your work will inspire further research and development in robot state estimation, and look forward to seeing its impact unfold in the field."
English
0
0
2
58
Gautam Kamath
Gautam Kamath@thegautamkamath·
Serious case of quality over quantity. I've seen PhD students with dozens of papers which I don't know anything about. But I've read these first two papers and they're both bangers. Calling this new one the best of the three is high praise and now I'm obligated to check it out.
Jeremy Cohen@deepcohen

@jasondeanlee @SebastienBubeck @tomgoldsteincs @zicokolter @atalwalkar This is the third, last, and best paper from my PhD. By some metrics, an ML PhD student who writes just three conference papers is "unproductive." But I wouldn't have had it any other way 😉 !

English
11
23
578
58.7K
Varun Agrawal
Varun Agrawal@varagrawal·
Woah!
Nico Bohlinger@NicoBohlinger

⚡️ Can one unified policy control 10 million different robots and zero-shot transfer to completely unseen robots, even humanoids? 🔗 Yes! Checkout our paper: arxiv.org/abs/2509.02815 If you are interested in massive multi-embodiment learning, come and chat with me at: - Today, WS Sim-to-Real Transfer for Humanoid Robots @Humanoids2025 - Oct 20th, WS Foundation Models for Robotic Design @IROS2025 - Oct 24th, WS Reconfigurable Modular Robots @IROS2025

English
0
0
1
229

Keşfet

@GeorgiaTech @MichaelAPosa @johnzhangx @OWW @fdellaert @keenanisalive @Slowbvrnerr_ @Compass_RP