Skip to content
ML Academy

SQL and Python

Interview question

  • seniorlevel
  • 30 minsuggested time
  • 3ML engineer, Data scientist, Data analyst

Whiteboard case. Design the nightly incremental pipeline that loads booking events from an operational system into an analytics table. Delivery is at-least-once, so duplicates occur; about 3% of events arrive one to five days late; schemas change occasionally; and analysts rerun history after bug fixes. How do you make it correct, idempotent and cheap?

Tests: Data pipeline design: idempotent partition writes, deduplication keys, late-arriving data, backfills, schema change handling and data-quality checks.

day 2day 6day 7day 19

Sign in to practise: answers are saved before model answers are shown.