RAG in practice: building AI on your own database

An AI service for finding coffee shops based on a proprietary database

You will learn RAG—the architecture behind serious AI products: instead of guessing an answer from the model’s general knowledge, the system first retrieves relevant data from your own database.

We’ll explore it through the example of a service for finding coffee shops in Belgrade. A user types “a quiet place with filter coffee near Kalemegdan,” and the service finds suitable venues by combining semantic search and keyword matching, taking location into account, and explaining why each place is a good fit.

You’ll learn how to apply RAG to your own use cases where you need AI-powered search over your data.

By the end of this case you'll have a working product and a set of reusable skills:

collecting data via external APIs: Google Places API

enrichment through web research: Parallel.ai

working with embeddings and vector databases: OpenAI, ChromaDB

building hybrid search: vector search plus keyword search with BM25

geographic search and query geocoding: Google Geocoding API

designing an interface on top of RAG: admin panel and user-facing website

How do I get access to the program?

This case is part of the GoPractice “Build with AI” subscription. You can start right now.

Learn more