Movie Recommender Model and Pipeline
Problem statement
How would you build a runnable movie recommendation pipeline during a live coding interview? You may use an AI coding assistant, but must review its code critically. The streaming service needs fast home-page rankings from viewing and impression logs, sparse ratings and title metadata. Fresh training and new users and titles make cold start important.
Choose labels and scaffold retrieval, ranking, training and evaluation.
Show hint
Before any code, decide what event in the logs means a user actually wanted a title, and what the tight response deadline implies about stages.
Relevant roles: ML Engineer, Applied Scientist and AI Engineer.
Related free problems
- ML Basics and Tuning for a Streaming RecommenderClassical MLMedium