Building an Event Recommendation Engine with Ahnlich and PostGIS

A developer describes how they built the recommendation system for Eventa, a distributed event ticketing platform for organizers and attendees. The platform runs eight backend services across NestJS and Go, communicating via gRPC, HTTP, Kafka, and RabbitMQ, with each service maintaining its own PostgreSQL database. The article focuses on the Discovery Service, which indexes published events and uses Ahnlich, a vector database for semantic search, to match events with attendee interests. Geographic searches and distance calculations are handled separately using PostGIS, a PostgreSQL extension. The developer explains that the project moved from a local Docker Compose setup to a VPS because their PC could no longer handle the containers.