Just published a case study on Smart Foot — an end-to-end embedded ML system I built from an empt...Just published a case study on Smart Foot — an end-to-end embedded ML system I built from an empt...
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Just published a case study on Smart Foot — an end-to-end embedded ML system I built from an empty PCB to a trained model.
It reads foot pressure in real time from a 16-channel FSR array, runs the whole pipeline in bare-metal C on ARM Cortex-M, and classifies gait with a neural network running on-device — no cloud, under 50ms.
The hardest part wasn't the firmware. It was collecting my own dataset: 1,500+ gait samples from 50 subjects, evaluated on a subject-independent split. That's why the 91%+ accuracy actually holds up on people the model never saw.
Full write-up in the comments 👇
Post image
Back to feed
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started