FastAPI service that ranks place recommendations using a scikit‑learn model
It solves the problem of ordering candidate places for a user in a geo‑social app by predicting relevance. The project generates synthetic data, builds contextual features, trains a scikit‑learn ranking model, and exposes FastAPI endpoints for ranking and explanation. It includes offline metrics (NDCG, Recall, MAP) and a comprehensive test suite, enabling easy integration or extension. Ideal for developers needing a reproducible, production‑ready recommendation backend.
View on GitHub →goghi48/ryden-ranker