Varad Pimpalkhute

@varad0309

RS @IFM_AI | Prev @Articul8_AI @AmazonScience @allen_ai | MS CS @UMassAmherst. Towards super intelligence, One Algorithm at a Time.

Sunnyvale, CA

Joined January 2021

721 Following

138 Followers

Replies Reposts Media

Pinned
Varad Pimpalkhute
@varad0309
Oct 22
Agentic RL is a whole new beast. Things broke. A lot. Pods crashed. Registries choked. Runs stalled. I spent the last few months in the trenches isolating failures and fixing bottlenecks. RL360 is what came out of it: millions of rollouts, scaled across multiple experts.
Institute of Foundation Models
@IFM_AI
Oct 22
We're open-sourcing RL360, the RL infrastructure behind K2 Horizon's six domain experts: 8,192 concurrent sandboxes, 150 steps in 7 days, ~1.3M rollouts. Long-horizon agentic RL runs thousands of agents in real sandboxes on tasks that take minutes. Some of those sandboxes crash
GIF
3
Varad Pimpalkhute
@varad0309
Oct 8
Interesting work from @Harman26Singh on measuring an agent's capacity to learn through experience and self-evolution.
Harman Singh @ ICML 🇰🇷🇰🇷
@Harman26Singh
Oct 8
🤖 How well do agents learn from experience? We study agents that amortize experience into reusable artifacts and introduce Agent Plasticity to measure how efficiently they improve on held-out environments. Top performer ≠ most efficient learner.
Varad Pimpalkhute
@varad0309
Sep 20
This is a really nice systems optimization. With router replay, the obvious optimization is usually to reduce how much routing state needs to be transferred. But there's a second dimension that matters just as much: when you transfer it. If routing decisions for the full
Rohan Pandey
@khoomeik
Sep 16
You saw the AI & science. Let's talk about the RL infra it took to build @periodiclabs Neon. To minimize training-inference mismatch in RL, SGLang captures inference's MoE routing decisions for each rollout and we "replay" them while training. In agentic (multiturn tool-use)
Varad Pimpalkhute
@varad0309
Sep 20
The thing that really sold me on UNO is that you don't need to maintain a separate draft model. That's always been one of the things holding me back from getting too excited about the usual speculative decoding regime. A separate drafter means another model to train/distill,
Institute of Foundation Models
@IFM_AI
Sep 17
Today’s LLMs still write like typewriters: one token at a time. This sequential process creates a hard inference bottleneck. We're introducing Uno, a diffusion-augmented LLM that delivers autoregressive quality at diffusion speed. It’s a lossless speedup method that accelerates
Varad Pimpalkhute
@varad0309
Sep 16
Massive respect! 🫡
Fuli Luo
@_LuoFuli
Sep 16
Nearly half a year of silence. We spent it studying one problem: how far RL can scale. MiMo-V2.6 is in the middle of its RL run right now. Three things we scaled: compute (~2B tokens per step, 1568 prompts × 16 rollouts, fully async), environments and harnesses (multi-task

Varad Pimpalkhute

Varad Pimpalkhute’s posts