As one of the founding engineers, I helped build an elderly-patient monitoring platform combining computer vision, IoT vitals, and medication analysis. Our prototype grew into SAP Homemade, now a product at SAP, and was a finalist for the Hasso-Plattner Innovation Award.
Highlights
- Used computer-vision Leigh lines to isolate wounds and color-ratio analysis to grade wound severity.
- Built CNNs with Royal Melbourne Hospital to detect pressure sores at 82% accuracy.
- Integrated sleep-cycle IoT devices (Dozee) to track patient health, plus medication analysis, at 99.99% accuracy.
- Prototype evolved into SAP Homemade, now a shipping SAP product, and was a Hasso-Plattner Innovation Award finalist.
Stack
PythonOpenCVKerasTensorFlowSAP UI5RedisGraphQL
Python
OpenCV
Keras
TensorFlow
SAP UI5
Redis
GraphQL