GenRec: Towards LLM-Native Recommendation at Netflix

Off-the-shelf LLMs make poor recommenders — they over-recommend popular titles, hallucinate items outside the catalog, and ignore business constraints. Netflix post-trained an internal foundation model on its own data and objectives to close that gap in production.

• LLM — A large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.
• hallucination — When a model states something false with full confidence — inventing facts, citations, or APIs that don't exist.

• blogMAPS: Netflix’s Multimodal Asset Personalization at Scale
• blogIn-House LLM Serving at Netflix
• blogGenPage: Towards End-to-End Generative Homepage Construction at Netflix

• articleNetflix PRS 2024 - Applying LLMs to Recommendation ExperiencesEugene Yan
• videoWhy LLM Recommenders Will Be AI's Biggest Consumer App — Devansh Tandon, MetaAI Engineer
• articleHow Generative Recommenders Are Redefining RecSys at ScaleElizabeth Goodman

Checking sign-in…

Loading comments…