Not pass or fail.
Know exactly where you lost points.

Write an answer to a real Data Engineering interview question. Get it scored against a rubric, with an exemplar response for every gap.

1,200+ questions · 54 topicsNo credit card required
1,200+
Practice questions
54
Topics covered
4 roles
DE · DS · DA · AE
AI-scored
Rubric feedback

How it works

Pick your topics
Choose a role (DE, DS, DA) and the topics you want to drill — SQL, Spark, dbt, system design, and more.
Answer and get scored
Write your answer in plain text. Our AI scores it against a rubric and shows exactly where you lost points.
Track your progress
See your mastery scores by topic, identify gaps, and use mock interview mode to simulate the real thing.

What you practice

System Design
Design data pipelines, warehouses, and architectures. Explain trade-offs with depth and clarity.
Ingest → Transform → Warehouse
SQL & Python
Write and review queries, explain window functions, optimise performance, and handle edge cases.
SELECT ... OVER (PARTITION BY dept)
Concepts & Theory
Answer questions on data modelling, orchestration, streaming, and cloud data platforms.
Modelling · Orchestration · Streaming

Stop guessing. Start knowing exactly where you stand.

Without polydomainai
Practicing alone with no idea if your answer was actually good.
Generic right-or-wrong feedback that doesn't say what to fix.
No sense of which topics you're actually weak in.
With polydomainai
AI-scored feedback against a real rubric, every time.
See exactly which criteria you missed, with an exemplar answer.
Mastery tracked by topic, so gaps are visible, not guessed at.

20 real DE interview questions

Click any question to sign up and see AI-scored rubric feedback on your answer.

Get started free
Spark

You have a Spark job that runs fine on 10M rows but OOMs on 500M rows. Walk me through how you'd diagnose and fix it.

Answer this
Spark

Explain the difference between SortMergeJoin and BroadcastHashJoin in Spark. When would you force one over the other?

Answer this
dbt

Your dbt model takes 45 minutes to run. How do you find the bottleneck and fix it?

Answer this
Ingestion

Design an incremental ingestion pipeline for a source table that has no updated_at column.

Answer this
Modeling

What's the difference between a fact table and a dimension table? Give a concrete example for an e-commerce company.

Answer this
SQL

Explain SCD Type 2. Write the SQL MERGE statement to implement it.

Answer this
Data Quality

You're seeing duplicate rows in your gold-layer table. Walk through the steps to identify the root cause.

Answer this
Contracts

What is a data contract? Who owns it, and what happens when a producer breaks it?

Answer this
Streaming

Compare Kafka delivery semantics: at-most-once vs at-least-once vs exactly-once. When does each matter?

Answer this
Orchestration

Your Airflow DAG has been running for 6 hours and is stuck. What do you check first?

Answer this
Streaming

Explain watermarks in Flink/Spark Streaming. What happens to late-arriving events?

Answer this
SQL

Write a SQL query to find the second-highest salary in each department without using LIMIT or TOP.

Answer this
Storage

What is partition pruning and how does Iceberg's hidden partitioning improve on Hive-style partitioning?

Answer this
Orchestration

You need to backfill 18 months of data. How do you do it safely without impacting production?

Answer this
Modeling

Explain the Medallion Architecture. What goes in Bronze, Silver, and Gold layers?

Answer this
SQL

What's the difference between RANK(), DENSE_RANK(), and ROW_NUMBER()? Give an example where they produce different results.

Answer this
Streaming

Your Kafka consumer lag is growing. What are the possible causes and how do you fix each?

Answer this
Ingestion

Explain Change Data Capture. What are the trade-offs between log-based CDC and query-based CDC?

Answer this
Spark

How does Spark's Adaptive Query Execution (AQE) work? What problems does it solve?

Answer this
Data Quality

Design a data quality test suite for a payments pipeline. What tests are blocking vs. warning?

Answer this

Simulate the real interview

Mock interview mode: pick a level (mid / senior / staff), go through timed rounds across topics, and get a full report with hiring verdict at the end.

Example DE rounds: Data Modelling → SQL → Python → System Design

Try a mock interview

Interview guides

Deep dives on what interviewers actually test.

All articles →

Built for data roles

100+ topics across Data Engineering, Data Science, and Data Analytics. Questions tagged by difficulty, topic, and subtopic — so you always know what you're working on.

Data EngineerData ScientistData AnalystAnalytics Engineer

Ready to level up?

Join for free. No credit card required. Start with 10 AI-scored practice sessions.