← back to work

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.

Building something similar?

I take on both freelance builds and full-time backend roles.

Get in touch