DokLink helps people in medical emergencies find and reserve hospital beds in real time. Users trigger an emergency, see nearby hospitals sorted by distance with live general and ICU bed counts, and reserve a bed with a countdown window sized to their travel time. I run the platform as CTO and sole developer: the React Native app, the Django REST backend, and the infrastructure it ships on.
Architecture
React Native + Expo client. Django REST Framework backend on PostgreSQL, with Cloudinary for media and scheduled jobs for reservation expiry. Deployed with Docker behind Nginx on a Linux server.
Features
Real-time bed availability with automatic release on cancel, admit, or expiry
Dynamic reservation windows based on travel distance
JWT authentication with OTP verification and Aadhaar validation
Payment integration with signature verification and transaction lifecycle handling
Challenges & Solutions
The booking flow is a textbook race condition: two users reserving the last ICU bed at the same moment. Solved with atomic transactions and row-level guarantees so bed counts can never go negative, plus automated expiry that returns unused reservations to the pool.
Tradeoffs
A single Django monolith instead of microservices. For a small team shipping fast in healthcare, one well-structured codebase with clear module boundaries beats a distributed system nobody has time to operate.
Performance
Booking-critical endpoints are kept lean: indexed geo-lookups for nearby hospitals and transaction-scoped writes for reservations. The app stays responsive on low-end Android devices, which is where emergencies actually happen.