Interview question asked to Data Scientists interviewing at Mozilla, Criteo, Teradata and other companies. Original question asked: How do you build automated sanity checks into an Airflow DAG to verify that the extraction layer pulled the exact expected number of records?.