Starlo is a production-grade hotel booking API built with Node.js, Express and TypeScript. It provides secure user management (signup/login/password reset + optional 2FA), advanced room discovery (filtering, sorting, pagination), booking lifecycle management, review aggregation, and admin tools. The stack was containerized with Docker, fronted by a load balancer, and enhanced with an in-memory caching layer to improve response times and availability.

Small-to-mid hotels and booking platforms need a resilient backend that handles search complexity, high read volume, concurrent bookings, and administrative reporting without becoming a maintenance burden. The core challenges were:

Complex search & filtering (many query combos) producing heavy DB load.

Concurrency risk during bookings (preventing double-book).

Need for real-time average rating updates and efficient review handling.

Requirement for maintainable, strongly-typed code (team handoff).

Need for resilience and horizontal scaling in production.

Node.js + Express + TypeScript — strong typing for safer refactors.

MongoDB — flexible document model for rooms & bookings.

Redis (or in-memory cache) — cache frequently-read queries (room lists, stats).

Docker + container orchestration behind a load balancer — for scaling and predictable deployments.

MVC structure + layered services — separation of concerns and testability.

Booking atomicity: booking endpoint uses optimistic checks and DB-level safeguards to avoid double-booking (document-level checks + unique constraints / transactions).

Advanced APIFeatures: query parser supports gt/lt/gte/lte, field selection, paging, and combined filters.

Review aggregation: when a review is created/updated/deleted, the room’s average rating is recalculated using incremental aggregation for constant-time cost.

Caching strategy: cache hot read endpoints (room listing, room-info) with TTL and cache invalidation on booking/review events.

TypeScript contracts: DTOs and controller/service boundaries reduce runtime errors and speed up maintenance.