Embedded Computer-Vision Sports System
Sports Performance Startup
The brief
A sports-performance startup wanted real-time computer vision built into a physical training aid — tracking fast-moving objects the instant they moved, with nothing on the critical path left waiting for the cloud.
We delivered the vision two ways over one shared pipeline. On a phone, GPU-accelerated computer vision on iOS and Android turned the athlete’s own device into the tracker — and alongside it, a purpose-built embedded IoT device could run that same real-time object tracking entirely on its own hardware, with no app required at all. A lightweight serverless AWS layer tied the two together for configuration, data and results.
