
Code @ Anthropic. Previously Amazon AGI Autonomy, @AdeptAILabs. Ph.D. @Berkeley_EECS. Undergrad @IITKanpur.
Indianapolis, IN
Joined April 2010
- Tweets73
- Following201
- Followers612
- Likes937
Rohan Bavishi retweeted

Anthropic@AnthropicAIFeb 28
United States
United StatesConnected via WebAccount-level information, not a live location or per-post device.
A statement on the comments from Secretary of War Pete Hegseth.
anthropic.com/news/statement…
Rohan Bavishi retweeted
Introducing Claude Opus 4.5: the best model in the world for coding, agents, and computer use.
Opus 4.5 is a step forward in what AI systems can do, and a preview of larger changes to how work gets done.
Rohan Bavishi retweeted
From spreadsheets to Airbnb listings, the ability to 'zoom out' using AI lets us quickly understand and act. It's not just about more data; it's about clearer, concise, actionable insights.
A look into what zooming out could look like from @Wattenberger and @AdeptAILabs
Rohan Bavishi retweeted
I’ve been working with @AdeptAILabs and we’ve made FlashAttention even faster for long sequences! For seqlen 8K, FlashAttention is now up to 2.7x faster than a standard PyTorch implementation even at small batch, making it easier to train better LMs with longer context 1/7
Announcing FlashAttention, a fast and memory-efficient attention algorithm with no approximation! 📣 w/ @realDanFu
By reducing GPU memory reads/writes, FlashAttention runs 2-4x faster & requires 5-20x less memory than PyTorch standard attention, & scales to seq. length 64K. 1/
Rohan Bavishi retweeted
This might have gotten lost, but this vid demonstrates voice input!
Complaints to me if you don’t like the voice in question :)
This quoted post is unavailable.
Rohan Bavishi retweeted
Train BERT on Smartphones 🤩 Announcing POET 📢 Find out how we train memory-hungry SOTA models on smartphones! #ICML2022 Thu 21 Jul 3:45 pm EDT at Room 327 🧵👇
Paper: arxiv.org/abs/2207.07697
Web: poet.cs.berkeley.edu/
Rohan Bavishi retweeted
So excited that our work "Gauss: Program Synthesis by Reasoning over Graphs", led by @code_monet, will be making its appearance at OOPSLA 2021 next week! Full paper: rbavishi.github.io/docs/oops… 1/5
Rohan Bavishi retweeted
I have openings for 2-3 postdocs to work on:
1. Whitebox fuzzing
2. Explore approaches to creating scientific software targeting distributed, heterogeneous, HPC platforms.
3. Fast data-engineering using program synthesis, AI, and semantic code search.
Email me if interested.
Rohan Bavishi retweeted
Our second paper, "VizSmith: Automated Visualization Synthesis by Mining Data-Science Notebooks", collaboration with @code_monet , @ShadajL , @koushik77 and Hiroaki Yoshida, also accepted to @ASE_conf !!
Rohan Bavishi retweeted
India's health system is on the verge of collapse
America can help.
My piece in @washingtonpost on how the world’s oldest democracy (the U.S.) can help the world's largest democracy (India).
It's time to act.
washingtonpost.com/opinions/…
Rohan Bavishi retweeted
I've shared this with many incoming PhDs picking a school, but I think it's worth sharing publicly. When you're choosing an advisor, talk to their students!! You're about to have a *lot* in common with them. If they're happy, solid chance you'll be happy. If they're miserable...
Rohan Bavishi retweeted
Folks on the academic job market: While there are many great guides about interviewing, probably not a lot cover "virtual" remote interviews, which you will soon be doing. Here are some thoughts based on my own experience winging it in 2020. Thread. 1/
Rohan Bavishi retweeted
Good news: The department has decided to completely remove the GRE (Graduate Record Examination) from consideration for graduate applicants for 2021 admission.
cc:
@Cal_Engineer
@BerkeleyDataSci
Rohan Bavishi retweeted
When I was and undergrad and junior grad student, I regularly burned myself out on deadlines. I actually work WAY harder now. And I'm able to do it because I learned to manage my energy better. Here are some lessons I learned.
Thread. 👇
Rohan Bavishi retweeted
#Chemistry #NobelPrizeLaureate Emmanuelle Charpentier: “The fact that two women won this Prize will hopefully show to young girls & young female scientists...that nothing is impossible”. 😊 The full video of yesterday's press conference: youtu.be/zZG8PhmLadk @maxplanckpress
Rohan Bavishi retweeted
Wow. In MIT/Harvard case, DHS has fully agreed to rescind the July 6 ICE guidance and July 7 FAQ, and rescind all implementing guidance!
So no TRO, because the government backed down completely.
Rohan Bavishi retweeted
.@SecPompeo .@DHS_Wolf
.@USDOL 30,000 Academic Workers at Univ of California urge you NOT to restrict or suspend F-1/OPT & H-1B visas. STOP the xenophobic bans on international students. Pls RT. @KamalaHarris @uaw2865 @uaw4121 @GWCUAW
uaw5810.org/wp-content/uploa…
Rohan Bavishi retweeted
Here is the link to the talk that I gave today. It's wonderful to know so many perspectives around superoptimization from different people. Feel free to pour in more suggestions. Thank you @johnregehr for helping me prepare for this talk!
Excellent talk by @jubitaneja on Improving Compiler Construction Using Formal Methods in the Berkeley Programming Systems Seminar. We had 46 participants tuning in---a record high! Check out the rest of the schedule at ps.berkeley.edu/seminar. youtube.com/watch?v=de8Ak0nY…


























