My work

Washudiwa

Built in: March, 2026 | Stack: TypeScript, Neon Postgres, TMDB

Finds Letterboxd users who love and hate the same films you do, then recommends what they loved and you missed. The thesis: one stranger who shares your two-star ratings is a better guide than ten thousand people who agree that a film is popular. Recommendations arrive with where to stream them. Say the name fast.

Architecture diagram:

FIG. 1 :: How it worked.
Text alternative

Your Letterboxd ratings flow into a shared pool alongside other users'. A taste-twin scorer requires both shared loves and shared hates to call someone your twin. Recommendations are what your twins rated highly and you have not seen, enriched with posters and streaming availability so you know where to watch tonight.