Altas Pelis

A fast movie recommendation system.

Check it out here: Altas Pelis

Altas Pelis is a content-based movie recommender system that uses Sentence Transformers to generate embeddings for each movie based on its details. It computes the cosine similarity matrix on-the-fly for each movie, and it does so very fast!

I've also published some of the code on GitHub.

Altas Pelis is a content-based movie recommendation system designed to find movies similar to the user's favorites. It uses Sentence Transformers to generate embeddings for each movie based on its details and calculates cosine similarity on-the-fly for recommendations.

The dataset includes over 10,000 movies, combining data from IMDb (directors, ratings, cast) and TMDB (overviews, posters, backdrops, keywords, production countries). Embeddings are stored as NumPy arrays; it is a very simple yet effective system.

The system is built with Next.js for the frontend and Flask for the backend, which provides endpoints for search, recommendations, and movie details.