A logistics company runs three pipelines (shipments hourly, inventory every 6 hours, analytics daily
A medium Pipeline Design mock interview question on DataDriven. Practice with AI-powered feedback, real code execution, and a hire/no-hire decision.
- Domain
- Pipeline Design
- Difficulty
- medium
Interview Prompt
A logistics company runs three pipelines (shipments hourly, inventory every 6 hours, analytics daily) glued by time offsets. On Black Friday a slow upstream broke the analytics output. Apply the entire L4 intermediate tier: (1) two orchestrators (one for the upstream cadence DAGs, one for the daily analytics) so each cadence runs separately; (2) replace any plain Snowflake mart between the upstream and analytics DAGs with a lakehouse format (Iceberg, Delta, or Hudi) so the analytics DAG fires on the asset's snapshot freshness rather than the clock; (3) add an observability_tool node monitoring the analytics output with an alert_destination so SLA misses route to on-call.
How This Interview Works
- Read the vague prompt (just like a real interview)
- Ask clarifying questions to the AI interviewer
- Write your pipeline design solution with real code execution
- Get instant feedback and a hire/no-hire decision