Anime-Recommender System

Built a production-grade 2-stage recommender system separating candidate retrieval and deep re-ranking. Designed using industry methodologies from YouTube’s seminal deep learning recommendation paper to optimize latency and relevance.