OPENING CHANNELS
home stack journey manifesto connect
// 005 โ€” talk to me

Find me in the
noise of the network.

I'm here to learn from builders and document. If you're working in AI, robotics, hardware, or any layer of the stack โ€” let's talk. DMs are open.

Let's build something.

The fastest way to learn is to talk to people already doing it. The second fastest is to build alongside them. I'm down for both.

๐•
DM me on Twitter / X@eshaantstack
โ†’
โœ‰
Send an emailhi@eshaant.dev
โ†’
{ }
Open an issue on GitHub@eshaantstack
โ†’
// note: I read every message. I reply to most. If you're building something interesting, expect a fast response โ€” I love swapping notes on architecture and debugging war stories.
~/connect.sh
$ whoami eshaant ยท 18 ยท hyderabad $ cat interests.txt โ†’ transformers, cuda, robotics โ†’ embodied ai, agents โ†’ hardware, first principles โ†’ writing, teaching, learning $ ./connect --method [1] tweet @ "@eshaantstack" [2] mail "hi@eshaant.dev" [3] github issue $ best_for โ†’ paper recommendations โ†’ build collabs โ†’ debugging horror stories โ†’ mentorship (i'll take it)
$
// everywhere else

Or find me around the web.

Same username everywhere. Usually. Sometimes I forget.

// ping me about

Things I want to talk about.

Topics where I'm actively learning and would love a conversation.

Transformer internalsattention, MoE, scaling laws
CUDA & GPU programmingkernels, memory, perf
Embodied AI / roboticssim2real, control, sensors
Hardware designverilog, fpga, chip basics
Build-in-public tacticswriting, threads, content
Open source collaborationPRs, issues, maintainers
Learning paths & curriculahow to study this stuff well
Cool side projectsshow me what you're building