Data Architect Interview Questions

Data Lake stores raw, unstructured or semi-structured data (e.g. files in S3, Parquet, JSON). You load first and transform later; it’s cheap and flexible. Data Warehouse stores structured, processed data optimized for analytics (e.g. Redshift, Snowflake, BigQuery). Schemas are defined; queries are fast. In practice, many companies use both: lake for raw and archival, warehouse for curated reporting and BI.

ETL (Extract, Transform, Load): Transform data before loading into the warehouse. Common in legacy, on-prem systems where the warehouse has limited compute. ELT (Extract, Load, Transform): Load raw data first, then transform inside the warehouse (e.g. with dbt, SQL). Preferred in modern cloud setups because the warehouse is scalable and you keep a single copy of the truth. Choose ELT when you have a powerful cloud warehouse and want flexibility to change transformations without re-ingesting.