Yamini Kamalpuria
Featured project
Queue Cure
76% of India's clinics run on paper token slips. Patients wait 2–3 hours with zero visibility — they don't know if they're 3rd or 30th in line. Receptionists manage everything from memory. One distraction and the queue breaks. There's no screen, no estimate, no way to check without walking up to the desk. Process Started by breaking the problem into two roles: receptionist and patient. The receptionist needs speed — adding a patient should take under 10 seconds, no friction. The patient needs visibility — current token, how many ahead, estimated wait. The core technical challenge was keeping both screens in sync instantly without either side refreshing. I chose Socket.IO for real-time bidirectional sync over a simple polling approach — any state change on the server broadcasts to all connected clients immediately. Built the queue as a server-side state machine so all logic lives in one place and both clients always see the same truth. Results Both screens sync live with zero page refresh. A receptionist can add a patient and assign a token in under 10 seconds. Wait time updates dynamically as the queue moves — calculated from real data, not a fixed guess. Deployed on Vercel + Railway with the live app accessible from any phone on the waiting room screen. Reflection I'd add a QR code on the receptionist screen that patients scan to open the waiting room view directly — removes the need to tell every patient the URL. I'd also persist queue state to a database so a browser refresh or server restart doesn't wipe the queue mid-day, which matters in a real clinic setting.