Design a guarded recommendation feed with explicit exposure data, candidate coverage, sparse-user fallbacks and temporal evaluation.
Learning path
- Recommendation objective and interaction log: define what success means
- Candidate retrieval: separate broad discovery from hard eligibility
- Implicit feedback: distinguish preference from what the system exposed
- Content and collaborative signals: give sparse learners a real fallback
- Ranking and re-ranking: balance relevance with coverage and curriculum rules
- Offline recommendation evaluation: replay catalog and learner state in time
- Project: build a guarded next-lesson recommender
Connected foundations
Use the earlier data and model boundaries as prerequisites.
Further paths
Practice
Continue into another subject: Graph Machine Learning Tutorial.