Method
Two-stage recommender
Candidate generation followed by a ranking model.
28 cases from 18 companies apply Two-stage recommender. 12 were checked in an online A/B test; 5 report results in production.
Typical tasks: Retrieve and rank from millions of items, Recommend to new users and new items, Balance relevance and diversity, Rank search results and 5 more. Companies: Pinterest, Meta, DoorDash, X5 Tech, Criteo, Whatnot and 12 more. Often combined with: A/B test, Learning to rank, Result diversification, Two-tower model, Gradient boosting and 50 more.



