Project
table-reservation
Restaurant table management system that reads real occupancy from security cameras instead of relying on staff to keep a board up to date. Pose-estimation models over RTSP streams detect seated customers and opens a table session, which then flows through waiter assignment and ordering. Reservations carry time-slot conflict detection, and staff work the floor through an AI-backed Telegram bot.
Pose estimation over RTSP camera streams
Automatic table sessions: detected, waiter assigned, ordered
Reservation lifecycle with time-slot conflict detection
Table CRUD with available/occupied/reserved status
Telegram AI agent for staff
JWT staff authentication