About

What is ITOHI?

ITOHI — Invent This, Invent That, Open Hardware Inventions. A place for projects that probably shouldn't exist yet.

Who am I?

My name is Andrius. My main strength is connecting hard technical dots to make things possible — the kind of things most people assume require a bigger team, more budget, or more time.

I have a physics degree in telecommunications and electronics, and I've spent the better part of two decades building at the intersection of hardware, firmware, and machine learning. The projects that interest me most live at the edge of what's technically feasible — where the answer requires knowing three unrelated fields at once.

Some examples of what that looks like in practice:

Sand art installation — a machine learning system that recognized when a participant started drawing pictures in sand and automatically saved clips for anyone to watch back.

Bus stop music box — software and firmware for a public installation that lets people play music of their choosing while they wait for a bus.

Visible light spectrometer — a portable spectrometer built around a Raspberry Pi, OV9281 camera, and TOSLINK fiber optic coupling. Designed for both reflectance and transmittance measurements.

Shahed drone detector — a neural net trained to detect and classify Shahed drone acoustic signatures. Runs inference in the browser. Multiple detection classes, outperforms DSP-only approaches on accuracy.

DARP — a patented decentralized autonomous routing protocol developed over five years. Peer-to-peer mesh networking without central infrastructure.

Currently prototyping a throwable WiFi CSI sensor for detecting life signs in enclosed spaces — designed to be deployed like a grenade, using the ESP32's onboard radio to detect human presence and movement without external sensors. Sub-$10 hardware, transmitting via ESP-NOW to a phone 10–20m away.

I also build AI helper workflows for technical consultancy through Mikonis Artisan MB.

How I work

Projects find me more often than I find them. I work best on hard technical problems where someone needs a 0→1 builder — the person who figures out whether the thing is even possible, then builds it.

I work well with independent researchers who need someone to actually construct the system, founders with unusual hardware problems, and anyone whose project keeps getting turned down by normal contractors because it's "too custom."

→ CV if you need the formal version → GitHub to see the work → Contact if something here overlaps with what you're building