Milan Sampath (@_0xMilan) header
Milan Sampath (@_0xMilan) profile picture

20, prev AI Engineer @aden_hq (YC W20) , building @cassettelabsin, prev @IND_stellar

Bangalore
Joined January 2023
Photo by Milan Sampath (@_0xMilan)Photo by Milan Sampath (@_0xMilan)Photo by Milan Sampath (@_0xMilan)Photo by Milan Sampath (@_0xMilan)Photo by Milan Sampath (@_0xMilan)Photo by Milan Sampath (@_0xMilan)Photo by Milan Sampath (@_0xMilan)Photo by Milan Sampath (@_0xMilan)Photo by Milan Sampath (@_0xMilan)Photo by Milan Sampath (@_0xMilan)
After securing the Best AI Blockchain position for Best AI x Blockchain project at the @Algorand Bangalore Edition, we are now heading to Delhi to participate in the AlgoDelhi Hackathon. Looking forward to building, learning, and representing on a larger stage. @mdmudassir0143
2
1
11
392
Sometimes I accidentally give respect to service based company majdoors, and it backfires on me
3
82
I finally quit my main job, it felt more like a side job to be honest, but its over, looking forward to more stuff in life
1
9
227
Day 2 of @EFDevcon , went to some of the discussions, met crypto folks. 💪
5
135
Day 1 of @EFDevcon 🫡🫡
2
15
353
finally gonna clean up some space on my mac
1
2
128
saw my first term sheet at 20. this might actually be getting serious.
2
113
taking the lab outside for a bit. cassette labs is coming to @EFDevcon . less talk, more experiments.
2
134
all about road to @EFDevcon @ethereumfndn , had an interactive sesh with a bunch of web3 folks @0xsaikarthik @diekneee
1
9
254
just @0xsaikarthik preaching @ethereumfndn @EFDevcon to the audience here at road to devcon bengaluru
5
2
19
624
not a good day until I see a $3400 daily deposit on my cex
7
137
I'm going to ETHGlobal Mumbai @ETHGlobal! Use my referral link below to earn a discount on your stake! Link: ethglob.al/IWWXES6W58
2
78
made this beauty, my wallpaper
Liftoff of Starship's first orbital flight
1
98
Just came across this interesting paper on Cache-to-Cache (C2C) communication between LLMs. Instead of making models communicate through text, it lets them share information directly through their KV-cache representations. The results are pretty interesting: better accuracy and significantly lower communication latency. Definitely an interesting direction for multi-LLM systems. check it out here: arxiv.org/pdf/2510.03215
2
6
368
took some time, but here it is. Ik it isn't a lot for a lot of people , but i had to work my ass off to afford this.
8
19
434
Studied Embedding Models in depth today Principal component analysis Singular value decomposition word2vec Bag of words Skip-gram BERT Super Interesting stuff if you wanna get started, checkout : aws.amazon.com/what-is/embed…
4
147