Problem Framing, Metrics, and Scale

~40 min read

How to scope a recommendation system: defining the problem, choosing the right success metric, establishing scale assumptions, and identifying the core challenges before touching any model design.

Problem Framing, Metrics, and Scale is a Pro topic

Sign in, then upgrade to Pro or Power to unlock this topic and the full AI Engineering curriculum.

Key points

  • The context (homepage vs. post-watch vs. search) changes the design — always clarify before designing
  • Choosing the right success metric is the highest-leverage design decision: watch time > CTR for video recommendation, but watch time alone can reward addictive content
  • Cold start (new users, new videos) and the filter bubble are first-class problems, not afterthoughts — name them in the framing phase