Automatic billing for truck parking
Cameras spot trucks entering and leaving a San Francisco parking yard, track how long each one stays, and bill for it automatically.
- Role
- Forward-Deployed AI Engineer (San Francisco Logistics Client via TechRAQ)
- Built with
- Python, OpenCV, PyTorch, FastAPI, PostgreSQL, Docker
Results
- Automated gate entry/exit tracking
- Real-time dwell-time billing ledger
- SF logistics MVP
In more detail
Forward-deployed logistics MVP for a San Francisco operator: computer-vision vehicle detection and entry/exit timestamp tracking pipeline that automates dwell-time calculation and dynamic billing for commercial truck parking facilities.
Key decisions
Multi-Camera Vehicle Detection & Gate Crossing Tracker
Processed live gate camera feeds using PyTorch/OpenCV object detection and directional line-crossing state machines to log commercial truck arrivals and departures under varying lighting conditions.
Automated Dwell-Time Reconciliation & Dynamic Billing API
Paired entry/exit events in PostgreSQL via FastAPI microservices to compute exact dwell durations, tiered rate schedules, and automated operator invoices.