Company
Meta
24 cases from Meta (Social, Adtech, Media streaming), each digested from a public talk or engineering post and linked to the original. 2 were checked in an online A/B test; 14 report results in production.
Tasks: Retrieve and rank from millions of items, Reduce ML cost, Recommend from what the user just did, Recommend to new users and new items, Optimize decisions and 11 more. Methods: Multi-task learning, Two-stage recommender, CTR / feedback prediction, Knowledge distillation, Two-tower model, Sequential recommendation and 36 more. Tools: PyTorch, FAISS, Llama and 3 more.