Design Top-K System — Section 6: Deep Dive — Ranking & Top-K Computation Pipeline
Subsystem Chosen: Ranking & Top-K computation
Why this matters: This is the core correctness + scalability bottleneck and a common Staff/Principal deep-dive topic.
1. Problem Restatement
Given:
- A stream of events
- Aggregated metrics per item
- A scope (surface, region, category)
- A time window
We need to:
- Compute a score per item
- Maintain a Top-K list
- Update incrementally
- Guarantee atomic visibility
Sign in to continue reading
The rest of this lesson is available with a free account. Signing in with Google or Microsoft is free.
Sign in to read the full lesson