trentenwen
I build small things that learn.
Engineer at the boundary where machine learning meets real hardware. Mostly TinyML, embedded systems, and the tools that make them work together. Always small.
- Based
- Somewhere with good coffee and a soldering station
- Currently
- Working on tinymlr · Embedded audio ML
- Available for
- Selected consulting, talks, and reviews
- Languages
- Rust, C/C++, Python, TypeScript, Lua
- Hardware
- STM32 (H7/F4/G0), ESP32, RP2040, RPi
- Reach me
- hello@trentenwen.com
I'm trentenwen. I build at the boundary where machine learning meets real hardware — quantized models on microcontrollers, ML-driven robotics, and the tools that make them usable.
Most of my time is spent on TinyML: pushing models down to kilobytes and microseconds, then wiring them to motors, microphones, and sensors that have to survive in the real world. Sometimes that means a 200-line runtime in Rust. Sometimes it means a 4 KB audio pipeline in C. Sometimes it means redesigning the same Kalman filter for the fourth time because the IMU is on fire.
I prefer small, sharp tools over big frameworks. A focused 200-line runtime will outlive a 2 MB dependency every time. Most of what I publish on this site is in that spirit: notes from the trenches of putting intelligence on devices that don't have much.
Outside of work, I keep a small bench of STM32s and ESP32s that I treat like a menagerie. Some of them make sounds. Some of them turn motors. Some of them exist only because I wanted to see if a thing was possible.