███████╗ █████╗ ███╗ ██╗████████╗██╗ ██╗ ██████╗ ███████╗██╗ ██╗ ██╔════╝██╔══██╗████╗ ██║╚══██╔══╝██║ ██║██╔═══██╗██╔════╝██║ ██║ ███████╗███████║██╔██╗██║ ██║ ███████║██║ ██║███████╗███████║ ╚════██║██╔══██║██║╚████║ ██║ ██╔══██║██║ ██║╚════██║██╔══██║ ███████║██║ ██║██║ ╚███║ ██║ ██║ ██║╚██████╔╝███████║██║ ██║ ╚══════╝╚═╝ ╚═╝╚═╝ ╚═══╝ ╚═╝ ╚═╝ ╚═╝ ╚═════╝ ╚══════╝╚═╝ ╚═╝ 🧠 Edge AI ⚙️ ShellOS 🌐 Full-Stack 🔬 Experimental ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
$ ./whoami --current-sprint
> 90-Day Polyglot Sprint 🚀
> Building: ESP32-LLM-Core + Edge-AI-SDK + Agent-Brain
> Goal: Make edge AI undeniable
> Velocity: 3 repos/week
> Status: ● LIVE & SHIPPING
$ neofetch --edge-ai
OS : ShellOS (ESP32)
Kernel : Bare-metal C + FreeRTOS
AI : LLM @ 20 tok/s offline
Power : < 100mW
Mission : Bring AI to every $4 chip
ShellOS — OS made Just For ESP32 Boards A full Linux-like shell OS running on bare-metal ESP32. Custom scheduler, file system, and shell — all under 100mW. 
|
ESP32-LLM-Core Full LLM inference at 20 tok/s on a $4 chip. No cloud. No latency. No privacy tradeoffs. Offline-first AI. 
|
Edge-AI-SDK pip install edge-ai. Abstracts TinyML chaos into 3-line inference calls. Works on any POSIX system + ESP32. 
|
Agent-Brain for ShellOS ShellOS-native autonomous agent. Perception → Reasoning → Action loop running entirely on-device. Zero API calls. 
|
 |
 |
 |
 |
 |
 |
"What if every $4 microcontroller could understand you? No cloud. No latency. No privacy tradeoffs."
© 2026 Santhosh | 🎯 QAI Happening | Full-Stack Edge Engineer