QQ'26

Traditional clinic queue systems are often manual, fragmented, and inefficient. Receptionists rely on verbal communication or paper tokens, causing delays, missed turns, and confusion in crowded waiting rooms. Patients often have no real-time visibility into queue progress, increasing frustration and uncertainty. Clinics also lack mechanisms for pause/resume handling, no-show detection, and live synchronization between staff and patients. The challenge was to build a low-latency, real-time queue management system that eliminates polling, minimizes manual coordination, and improves the waiting I approached the problem by mapping the clinic workflow into distinct real-time events: patient registration, token progression, queue pausing, and no-show handling. Instead of traditional polling-based architectures, I designed the system around Supabase Realtime to enable instant event propagation using postgres_changes subscriptions. The frontend was built with React + Vite + TypeScript for fast UI rendering and modular architecture, while Zustand handled state synchronization efficiently across components. I structured the application into two interfaces: Receptionist (control layer) and Waiting Room (display layer). Early iterations considered periodic polling for updates, but this introduced unnecessary load and latency. I shifted to a zero-polling architecture, which significantly The final system achieved real-time synchronization with near-instant queue updates across all connected clients without polling. Manual coordination steps between receptionist and patients were reduced significantly, streamlining clinic operations. The system supports dynamic queue pausing, automatic no-show handling, and live "Now Serving" updates, improving patient transparency and reducing missed turns. The modular architecture also improved maintainability and scalability for future extensions like multi-doctor queues and analytics dashboards. Given more time, I would expand the system to support multi-counter and multi-doctor queue management, allowing parallel consultations. I would also introduce predictive wait-time estimation using historical queue data and AI-based no-show prediction. Another improvement would be role-based authentication for different clinic staff and a mobile-friendly patient notification system for token alerts. These additions would make the system more scalable for larger healthcare environments.