Twitxy Saved

Bidipta Sarkar

@bidiptas13

EGGROLL Author | PhD Student at @flair_ox and @whi_rl | Stanford BS CS '24 @StanfordAILab | Ig @bidiptas13
bsarkar321.github.io Joined September 2021
121 Following 1.1K Followers 297 Posts

Latest posts 5

Pinned post

Bidipta Sarkar @bidiptas13
Introducing 🥚EGGROLL 🥚(Evolution Guided General Optimization via Low-rank Learning)! 🚀 Scaling backprop-free Evolution Strategies (ES) for billion-parameter models at large population sizes

⚡100x Training Throughput
🎯Fast Convergence
🔢Pure Int8 Pretraining of RNN LLMs
19 replies 153 reposts 998 likes 312K views View on X
Bidipta Sarkar @bidiptas13
Build and share a project using EGGROLL this month!
Daria Gneusheva Verified account @DGneusheva Aug 3
What if we could train LLM at the speed of inference with less compute?

We had a pleasure of hosting @bidiptas13, co-inventor of EGGROLL algorithm, to talk about the potential of evolutions strategies.

Build something creative with EGGROLL and share on X, tagging @outlierlabshq and @bidiptas13 before 26th of August.

Get featured by Bidipta and win a dinner with the Eggroll team!

Link to EGGROLL tutorial in the comments.
View quoted post on X
Bidipta Sarkar @bidiptas13
BOLD admins didn’t let me select the better cover image because “not furry enough” :(
Another #ICML2026 BOLD paper, led by @bidiptas13, Mattie Fellows and @JuanDuquevan. We introduce EGGROLL, a novel general-purpose machine learning algorithm that provides a hundredfold increase in training speed over naïve evolution strategies.

EGGROLL practically eliminates the barrier between inference and training, allowing us to easily fine-tune LLMs for reasoning or train new architectures from scratch.

Check the paper at: arxiv.org/pdf/2511.16652
View quoted post on X
4 replies 6 reposts 43 likes 7.3K views View on X