5 patterns. 7 datasets. 20 challenges.

The 20 challenges in this pack collapse to 5 recurring patterns. Learn the datasets once, then the patterns once, and you'll recognize the shape of every interview question — instead of solving each one cold.

How to use this playbook

  1. Start with Datasets — get the table shapes and FKs in your head once.
  2. When stuck on a challenge, jump to its pattern page from the Deep dive tab — read the skeleton + the trap.
  3. Each pattern page is ~1 scroll: TL;DR, tables used, copy-pasteable SQL skeleton, the trap, related challenges.

Reference

The Senior Loop

Framework

The 5-step mental loop every SQL interview question fits. Read once to stop solving each challenge cold — the scaffold transfers.

Datasets

7 tables

The pack's tables themselves — ERD, columns, common filters and joins, plus the cross-cutting gotchas every Netflix query has to handle.

Data Map

Interactive

Netflix's real-world data & analytics architecture — sourced from public tech-blog posts. Read it once for interview-prep priming.

Patterns

5 families · covers all 20 challenges

Engagement Metrics

5 challenges

Aggregate playback with caps, percentiles, and time windows — without falling for rewatch inflation.

Funnels & Sessions

3 challenges

Stitch impression → detail → play with attribution windows; derive sessions from event streams.

Workflow Ops

4 challenges

Pair stage_started ↔ stage_completed, detect in-flight via anti-join, count rework loops.

Retention & Cohorts

3 challenges

Gaps-and-islands streaks, monthly signup cohorts, multi-signal churn risk — censoring the incomplete.

Experimentation

4 challenges

Variant joins, SRM checks, ITT vs TOT, guardrail metrics — the experiment platform questions.

Ads Measurement

3 challenges

Reach & frequency, last-touch attribution windows, CTV quartile completion — the 1P/3P measurement vocabulary.