Full-Stack · ML
MatchFlix
A swipe-based movie discovery app where groups match on what to watch together, with LLM-powered recommendations from swipe behavior.
- Role
- Full-stack developer
- When
- 2025
- Stack
- Node.jsExpressPostgreSQLDockerTMDB APIGeminiOpenAI
TL;DR
- A full-stack movie discovery app. Swipe right or left on movies to build a personalized watchlist.
- Swipe Sessions let a group swipe together and match on the movies everyone wants to watch.
- Generative AI (Gemini, OpenAI) turns swipe history into personalized recommendations.
The problem
Picking a movie alone is easy. Picking one with friends turns into twenty minutes of scrolling and "I don't mind, you choose." MatchFlix turns that into a quick, low-stakes game.
How it works
- Swipe to rate. Movies come from the TMDB API. Likes and dislikes are stored per user and feed a personal watchlist.
- Swipe Sessions. Several users join a shared session and swipe on the same pool. Matching logic in PostgreSQL finds the movies the group agrees on.
- AI recommendations. A user's swipe behavior is summarized and sent to Gemini / OpenAI, which suggests what to watch next.
- API and deployment. A RESTful Express API with authentication, containerized with Docker so development and deployment behave the same.
Stack
Node.js · Express · PostgreSQL · Docker · TMDB API · Gemini API · OpenAI API
What I learned
Building the group feature meant thinking about shared state (many users writing to one session at once) rather than a single user's data, and designing the schema around that from the start.
Next project
Facial Recognition Attendance