ShelfEye
Retail attention intelligence โ tracks where shoppers look using head-pose estimation through existing CCTV cameras. 9-zone heatmap dashboard, live dwell-time analytics. No hardware. No privacy violations.
I build software-only AI systems that ship to production โ from head-pose estimation pipelines that replace $40k hardware to event-driven security cameras with sub-second alerting. Strong in Python, OpenCV, MediaPipe, Flask, React and AWS. Prefer deploying real products over writing academic reports.
Production-grade systems built from first principles.
Retail attention intelligence โ tracks where shoppers look using head-pose estimation through existing CCTV cameras. 9-zone heatmap dashboard, live dwell-time analytics. No hardware. No privacy violations.
Event-based security camera system with real-time motion detection. Records video clips only on trigger events and dispatches email alerts with footage in under 1 second. Zero cloud dependency.
Native desktop GUI for institutional student records. Full CRUD operations, persistent SQLite backend, and table model capable of handling 10k+ records with sub-50ms query times. No subscription required.
Languages, frameworks and tooling drawn directly from shipped projects.
Open to internships, collabs, and interesting problems. Message goes straight to my inbox.