100% ON-DEVICE & HARDWARE AI SYSTEMS
Building software
that runs where
the hardware is.
On-device AI, embedded systems, Android/Linux tooling, hardware experiments and security research — built from constrained machines.
Shipped systems
Systems
Eleven systems are shipped and thirty-three repositories are public.
Repositories
Thirty-three public repositories.
Loading repository list
Signal
Software, models, hardware and security are one job seen from different layers.
Device stack · local execution
The whole stack.
One device.
A phone becomes the workstation: Android hosts Termux, local models run beside the tools, and the same setup reaches real hardware.
- 01Android
- 02Termux + Linux
- 03Local AI
- 04ESP32 / hardware
FIG. 01 Phone → local runtime → hardware
Field notes
About
Built on Android + Termux · Shipping since Eight years
Krish builds systems that run on hardware nobody markets.
The hardware here is usually a phone.
Eleven systems are shipped and thirty-three repositories are public.
The toolchain is Termux, Node, Python and proot.
Models run on the device.
The microcontroller on the bench is an ESP32, reached over USB-OTG or BLE.
Software, models, hardware and security are one job seen from different layers.
Open to frontend roles, systems work and on-device AI projects.
Reach him at github.com/Axeb2b.
Bench record
The lab
20 pieces
Contact
Contact
Open to frontend roles, systems work and on-device AI projects.