AI Matchmaking Platform
Love2Knot
Backend and AI recommendation engine for a matchmaking platform, matching people using compatibility scoring rather than simple filters.
role
Backend Developer
period
Feb 2024 — Aug 2024
client
Tech Tank Innovations Pvt Ltd. (Delhi)
backend performance gain
35-40%
LaravelPythonReact.jsRedisMySQL
the challenge
Basic filters (age, location) weren't producing matches people actually wanted. The product needed a recommendation layer that scored real compatibility, without slowing the app down as the user base grew.
the approach
- Kept Laravel as the system of record for auth, profiles, and the product surface, and built the scoring logic as a separate Python service with a well-defined internal API contract.
- Designed a Python-powered AI Match Recommendation Engine using compatibility scoring and intelligent matching algorithms to rank candidates per user.
- Cached scores that didn't need per-request recalculation in Redis, and moved heavier scoring work into queue-based background processing.
- Integrated payment gateways, cloud services, and webhook-based third-party APIs for secure transactions and real-time communication.
results
- Backend performance improved 35–40% through Redis caching, queue-based processing, and database optimization.
- Recommendation engine shipped as an isolated, independently scalable service rather than a monolith rewrite.
- Secure payment and communication flows integrated end-to-end.