Loading...
Loading...

I build across hardware, software, and ML—PCBs to CLI tools, firmware to desktop apps, model training to on-device inference. Owning the whole stack is why elektroThing exists.
The interesting problems live at the intersections. A terminal roguelike that reads your git history. An AI coding assistant that runs entirely offline. A desktop Tamagotchi built in Rust. A social app where you send procedurally-generated art to friends. On the hardware side: wearable trackers with on-device ML, audio DSP platforms, USB-C power analyzers. I go where the problem takes me—TypeScript or C++, Tauri or KiCad, neural network optimization or PCB layout.
KiCad for hardware, C/C++ for firmware, Python and PyTorch for model training, TensorFlow Lite and Edge Impulse for edge deployment, Next.js and Supabase for web apps. I jump between domains because that's where the leverage is—understanding the hardware constraints shapes how I design the model, and understanding the math makes efficient DSP possible. Most of what I ship comes with docs, schematics, and enough context that you can actually build on it.
I kept building tools for my own projects and realized others might find them useful too. Having broad expertise lets me own the whole stack—and open-sourcing everything means you can learn from it, modify it, or just buy one ready-made. The community taught me through forum posts and GitHub repos; this is me paying it forward.
Got a project that doesn't fit neatly into one discipline? Those are my favorite kind. Let's talk.
Get in Touch(opens in new tab)