Shreeyash Pandey
@b0jle
systems dev @Vicharak_In, ml compilers, FPGAs, EDA, GPUs and fun
our talk 'No-ISA is the best ISA' is now live - check out the video where we break down our FPGA ideas at @Vicharak_In
sada dosa with butter in the centre
the Milky Way is not only flying at 600 km/s, but also “flapping its wings” like a butterfly🦋
the Milky Way is not only flying at 600 km/s, but also “flapping its wings” like a butterfly🦋
Pls rt fot good karma. Had so much fun cooking this with Sachin 🥳🥳
Excited to share Aryabhatta 1.0, our leading model that scores 90.2% on JEE Mains, outperforming frontier models like o4 mini and Gemini Flash 2.5 Trained by us at @AthenaAgentRL , in collaboration with @physics__wallah, using custom RLVR training on 130K+ curated JEE problems…
Python is an amazing language. It needs more work on its type system, but no language is close for the ability to express raw ideas with minimal overhead.
i discovered astral and tinygrad at approximately the same time. changed my views towards python completely
世の中には、ビジネス的には凄い価値なのにアカデミック的に価値が無いので reject された大学の研究成果とか、 アカデミック的には大発見なのにビジネス価値が無いとして発表許可の裁可が下りなかった企業内の研究とか、産学の狭間でそういうのがそれなりにありそうな気もしなくもない。
#Rockchip RK182x co-processor for LLM. LLM on chip is on the way. #RKDC2025
DeepTech != SlowTech. If we wait for perfect: - we overfit to theories, not usage - we build things nobody asked for - we ship slower, and probably heavier It’s easy to hide behind "deep tech" as an excuse to move slow.
Our bottleneck is no longer writing code, but reviewing code
babe wake up, new app dropped you can make a breakout board for ANY component breakout.circuitboard.club
messing with the rk3399, i decided to benchmark memcpy times. there's a sudden spike near the 1000 KiB mark, which perhaps could be explained by the L2 cache being 1MiB.
Every year at DAC, a session is held on discussing open source EDA: Open Source EDA Birds of a Feather …urce-eda-birds-of-a-feather.github.io