Deloitte Data Engineer Interview Guide

Candidates describe the Deloitte Data Engineer interview as practical and tied to the stack rather than algorithm-heavy. SQL, PySpark, Python and cloud pipeline scenarios come up most often. Several people were asked highest-salary SQL queries in some form, from the top three salaries to the highest salary per department and the second-highest salary. Others reported window functions such as rank versus dense_rank, duplicate records, MERGE statements and slowly changing dimension Type 2.

Spark questions go past syntax. One candidate had to diagnose a slow Spark job. Another was asked how to join two large tables and handle the small-file problem. A third was asked how Spark breaks work into jobs, stages and partitions. The cloud platform depends on the team: reports mention GCP with Airflow and Pub/Sub, Azure, AWS, Snowflake, and Databricks with Delta tables.

Project discussion is a recurring thread. Interviewers pressed candidates on why they chose an architecture, what each component does, and how they would monitor a pipeline, add data quality checks and handle failures. Specific stories from your own work serve you better than definitions. One candidate had to formulate SQL verbally without seeing the schema, so practice stating your assumptions aloud.

Most reports describe two or three rounds, usually two technical interviews followed by a managerial or HR conversation. One candidate instead had a phone screen and then a presentation round prepared from a briefing pack, with deliberately confrontational consultancy-style questions. A few were asked basic Python, such as reversing a string, or what it means to normalize data. A couple of candidates mentioned slow or irregular communication, so be prepared to follow up on your status.