FedEx Dataworks Data Engineer Interview Questions (2026)

A **Data Lake** is a centralized repository storing vast amounts of raw, multi-structured data (structured, semi-structured, unstructured) in its native format, applying schema-on-read for flexibility. A **Data Warehouse**, conversely, is a highly structured and curated repository optimized for analytical queries, enforcing schema-on-write for data consistency and performance. ### Mechanics and Use Cases Data Lakes, typically built on low-cost object storage like AWS S3 or Azure Data Lake Storage (ADLS), ingest data as-is. This makes them ideal for exploratory analytics, machine learning training, and long-term archival of diverse datasets (e.g., IoT sensor data, web server logs, social media feeds). Tools like Apache Spark, Presto, or Trino apply a schema dynamically at query time, allowing for evolving data needs without costly upfront schema design. Data Warehouses (e.g., Snowflake, Amazon Redshift, Google BigQuery) are designed for business intelligence (BI) and reporting. Data undergoes strict ETL/ELT processes to conform to a predefined schema (e.g., star or snowflake schema) upon ingestion. This schema-on-write approach ensures data quality and enables highly optimized, performant SQL queries for dashboards and reports, often leveraging columnar storage and advanced indexing. ```sql -- Data Warehouse: Schema-on-Write example CREATE TABLE sales.daily_transactions ( transaction_id VARCHAR(50) PRIMARY KEY, product_sku VARCHAR(20) NOT NULL, sale_date DATE NOT NULL, quantity INT NOT NULL, amount DECIMAL(10, 2) NOT NULL ); -- Data loaded into this table *must* adhere to this structure. ``` ### Key Trade-offs The primary trade-off lies in flexibility versus governance and performance. Lakes offer unparalleled flexibility and cost-effectiveness for storing all data, but require significant engineering effort for data quality, metadata management, and security to prevent a "data swamp." Warehouses provide high performance, strong data governance, and ease of use for BI, but can be expensive at petabyte scale, especially with high compute demands. Modern cloud data warehouses, like Snowflake, separate compute and storage, offering better scalability and cost efficiency than traditional systems. In the interview, also mention the **Lakehouse architecture** (e.g., Delta Lake, Apache Iceberg), which merges the flexibility of data lakes with the ACID transactions, schema enforcement, and data quality features typically found in data warehouses.