Hasith Vattikuti

Visiting researcher at TrainloopAI, Incoming CS PhD @Caltech, Previously @UTAustin,

[Simran Arora @simran_s_arora]

AI compute and inference are increasingly $$$. How can we change the unit economics of AI to improve accessibility? It's been fun working with @prlnet to release the first model endpoint that simultaneously generates tokens **and** a digital asset that can subsidize inference! 🪙 Check it out, links below 🚀

[Hasith Vattikuti @hasith_v]

Loved working on this project, super proud of it! Thank you @AwesomeBao, @bialjail, and @wgilpin0 for being such a great team that I learned so much from.

Excited to share our work at ICML!

[William Gilpin @wgilpin0]

Our ICML spotlight paper discovers universal redundancies in time series foundation models: the middle layers of many models can be removed without sacrificing performance 1/

[Jackson Stokes @jackson_stokes]

We trained LoRA adapters of different ranks to understand training dynamics, finding that adapters for GSM8k live in a surprisingly vast, low-rank solution space.

This hints that some model skills are easy to learn, and training is more forgiving than we think. @hasith_v 1/6 🧵

[William Gilpin @wgilpin0]

Congratulations to @BaigYasa on his well-deserved selection as a PD Soros fellow!

[Jackson Stokes @jackson_stokes]

We post-trained MedGemma to be SoTA in visual medicine ddx, outperforming Opus 4.6, Gemini 3.1 and GPT-5.4 while running at ~1/30th the cost. @getnolla Part 1 - improving visual reasoning 🧵1/6

[Hasith Vattikuti @hasith_v]

@jxmnop This is cool, I've always been slightly uncomfortable with treating *everything* unlabeled as a negative.

I wonder if using LLMs to produce rankings (even a somewhat noisy one) would be better than a binary classification. Perhaps we can weight according to rank in softmax?

[William Gilpin @wgilpin0]

How do time series foundation models forecast unseen dynamical systems? In new experiments, we find that small transformers learn to approximate transfer operators in-context. (1/N)
https://t.co/6YuLr8QuJD

[dhruva @dhruvakarkada]

Soooo proud of this one! I'll make a post w more details shortly

[Hasith Vattikuti @hasith_v]

@jxmnop Will code be released? Interested in playing around with this

[Hasith Vattikuti @hasith_v]

@khoomeik @LEGO_Group ASML shut down all talks of a collab in fears of trade secrets being leaked in the build

[Yasa Baig @BaigYasa]

Great to see high quality software dev in comp bio. It still amazes me how much of computational biology is based on single-thread processing of large .txt files with minimal application-specific-optimization.

[Hasith Vattikuti @hasith_v]

@sidbing Share a list of your favorites!

[Hasith Vattikuti @hasith_v]

@mattarderne @karpathy See the section titled “The LLaDa Algorithm” in my blog post

[Hasith Vattikuti @hasith_v]

@karpathy https://t.co/DiZH86Wlnn

[Hasith Vattikuti @hasith_v]

@karpathy I actually hacked nanogpt sometime ago to become a diffusion llm. Results were pretty decent on shakespeare with character-level tokenization.

Honestly was just surprised it even learned to spell words and pick up on basic grammar.

Link in reply

[Hasith Vattikuti @hasith_v]

@materzynska @AIatMeta Very interested in diffusion models and social AI. Would love to talk with you. You can see more about me on my blog: https://t.co/BUGkLLEquN

[Hasith Vattikuti @hasith_v]

@a16z @LiamFedus @LiamFedus what are yalls methods to verify what the LLMs are discovering? How do you make sure it’s ‘understanding’ current physics correctly?

I have lots of thoughts on this as a physics student doing AI research if you want to chat

[Hasith Vattikuti @hasith_v]

@khoomeik @periodiclabs @LiamFedus Very excited to see where periodic will go next! Extremely bullish on trying to get tangible alpha from AI models in natural sciences--it really plays to my background of first doing physics research and then doing AI research

[Hasith Vattikuti @hasith_v]

@CFGeek Yes it is, happy to discuss and get feedback. All is welcome

[Hasith Vattikuti @hasith_v]

@CFGeek To be fair, I also think it will be hard to get it to work, and it might not even. But the negative result plus the rl env will leave us things to learn from.

Cause I’m pretty confident that LLMs will be using internal reasoning techniques only a few years down the line.