Data Engineer interview questions

Tell me about a time you owned a data quality problem that other teams depended on.

From the Google guide · Practice it in a mock interview · Draft my answer from my resume

Tell me about a time you owned a data pipeline that failed in production. What did you do and what changed afterwards?

From the Amazon guide · Practice it in a mock interview · Draft my answer from my resume

Describe a time you disagreed with a stakeholder about a data model and how it was resolved.

From the Amazon guide · Practice it in a mock interview · Draft my answer from my resume

How would you make a daily job safe to rerun after a partial failure?

From the Meta guide · Practice it in a mock interview · Draft my answer from my resume

Tell me about a data model you designed that needed to change after launch.

From the Meta guide · Practice it in a mock interview · Draft my answer from my resume

Tell me about a data model you designed that had to change after launch. What did you get wrong?

From the Microsoft guide · Practice it in a mock interview · Draft my answer from my resume

Write a query that finds the first failure event for each device within a rolling window and explain how you would test it.

From the NVIDIA guide · Practice it in a mock interview · Draft my answer from my resume

Tell me about a data model that had to change after launch and what you would design differently.

From the NVIDIA guide · Practice it in a mock interview · Draft my answer from my resume

How would you make a nightly job safe to rerun after a partial failure?

From the LinkedIn guide · Practice it in a mock interview · Draft my answer from my resume

Tell me about a data model you designed that had to change after launch.

From the LinkedIn guide · Practice it in a mock interview · Draft my answer from my resume

Write a query that finds the first failure for each device within a window and explain how you would test it.

From the Apple guide · Practice it in a mock interview · Draft my answer from my resume

Why do you want to work on this team's data problems in particular?

From the Apple guide · Practice it in a mock interview · Draft my answer from my resume

Tell me about a data model that had to change after launch and what you would design differently.

From the Apple guide · Practice it in a mock interview · Draft my answer from my resume

Tell me about a data model that had to change after launch and what you would design differently.

From the Tesla guide · Practice it in a mock interview · Draft my answer from my resume

Why do you want to work on this team's data problems?

From the Tesla guide · Practice it in a mock interview · Draft my answer from my resume

Write a query that finds accounts whose settlement totals disagree with their transaction totals and explain how you would test it.

From the Stripe guide · Practice it in a mock interview · Draft my answer from my resume

Tell me about a data quality problem you owned that other teams depended on.

From the Stripe guide · Practice it in a mock interview · Draft my answer from my resume

Tell me about a data model that had to change after launch and what you would design differently.

From the Intel guide · Practice it in a mock interview · Draft my answer from my resume

Write a query that identifies customers whose orders were shipped late in consecutive months and explain how you would validate it.

From the Amazon guide · Practice it in a mock interview · Draft my answer from my resume

Write a query that finds users who were active on consecutive days and explain how you would test it.

From the Meta guide · Practice it in a mock interview · Draft my answer from my resume

Write a query that finds customers whose spend increased month over month and explain how you would validate the result.

From the Microsoft guide · Practice it in a mock interview · Draft my answer from my resume

Write a query that finds members who were active on consecutive days and explain how you would test it.

From the LinkedIn guide · Practice it in a mock interview · Draft my answer from my resume

Write a query that finds the first fault event per vehicle within a rolling window and explain how you would test it.

From the Tesla guide · Practice it in a mock interview · Draft my answer from my resume

Write a query that finds the first failing measurement for each unit within a window and explain how you would test it.

From the Intel guide · Practice it in a mock interview · Draft my answer from my resume

What does QUALIFY do, and what would you write without it?

Practice it in a mock interview · Draft my answer from my resume

What is a slowly changing dimension, and how do you implement type two?

Practice it in a mock interview · Draft my answer from my resume

How would you design a pipeline that ingests clickstream events and makes them queryable within minutes?

From the Google guide · Practice it in a mock interview · Draft my answer from my resume

Design a system that ingests order events from several regions and makes them available for daily reporting and near real-time alerts.

From the Amazon guide · Practice it in a mock interview · Draft my answer from my resume

Design a pipeline that turns raw app events into daily engagement metrics and explain how you would handle events that arrive a day late.

From the Meta guide · Practice it in a mock interview · Draft my answer from my resume

A metric dropped after a schema change upstream. How do you find out whether the drop is real?

From the Meta guide · Practice it in a mock interview · Draft my answer from my resume

Design a pipeline that loads daily sales files from several regions into one reporting table and handles a late region without blocking the others.

From the Microsoft guide · Practice it in a mock interview · Draft my answer from my resume

A downstream dashboard shows numbers that do not match the source system. How do you find where the difference was introduced?

From the Microsoft guide · Practice it in a mock interview · Draft my answer from my resume

When would you choose a streaming design over a scheduled batch, and what does it cost you?

From the Microsoft guide · Practice it in a mock interview · Draft my answer from my resume

Design a pipeline that collects telemetry from thousands of devices and makes it queryable for both daily reports and ad hoc analysis.

From the NVIDIA guide · Practice it in a mock interview · Draft my answer from my resume

A downstream team says yesterday's totals are wrong. How do you find where the pipeline introduced the error?

From the NVIDIA guide · Practice it in a mock interview · Draft my answer from my resume

When would you choose a streaming design over scheduled batch, and what does it cost in complexity?

From the NVIDIA guide · Practice it in a mock interview · Draft my answer from my resume

Design a pipeline that turns member activity events into daily engagement metrics and handles events that arrive a day late.

From the LinkedIn guide · Practice it in a mock interview · Draft my answer from my resume

A dashboard shows a drop that the source system does not. How do you find where the difference was introduced?

From the LinkedIn guide · Practice it in a mock interview · Draft my answer from my resume

Design a pipeline that ingests device diagnostics from millions of devices and makes them queryable for daily reports and privacy-safe analysis.

From the Apple guide · Practice it in a mock interview · Draft my answer from my resume

A daily metric dropped after an upstream schema change. How do you find out whether the drop is real?

From the Apple guide · Practice it in a mock interview · Draft my answer from my resume

Design a pipeline that ingests telemetry from a large vehicle fleet and makes it queryable for daily reports and near real-time alerts.

From the Tesla guide · Practice it in a mock interview · Draft my answer from my resume

A factory dashboard shows totals that do not match the source system. How do you find where the difference was introduced?

From the Tesla guide · Practice it in a mock interview · Draft my answer from my resume

Design a pipeline that reconciles payment events from several sources into a daily ledger that must never double count.

From the Stripe guide · Practice it in a mock interview · Draft my answer from my resume

Here is a pipeline you have never seen and a failing job. Find the cause.

From the Stripe guide · Practice it in a mock interview · Draft my answer from my resume

Design a pipeline that collects test data from many manufacturing lines and makes it queryable for daily yield reports and ad hoc analysis.

From the Intel guide · Practice it in a mock interview · Draft my answer from my resume

A daily aggregate disagrees with the source system. How do you find where the pipeline introduced the difference?

From the Intel guide · Practice it in a mock interview · Draft my answer from my resume

Design the tables for an online marketplace's order history. What is the grain of your fact table?

Practice it in a mock interview · Draft my answer from my resume

Explain a star schema, and when you would deliberately denormalise further.

Practice it in a mock interview · Draft my answer from my resume

How do you make a pipeline idempotent, and why does it matter for reruns?

Practice it in a mock interview · Draft my answer from my resume

How would you detect that a pipeline succeeded but produced wrong data?

Practice it in a mock interview · Draft my answer from my resume

Walk me through a data model for a subscription business and explain how you would handle plan changes over time.

From the Google guide · Practice it in a mock interview · Draft my answer from my resume

A nightly job started producing duplicate rows. How do you find the cause and keep it from happening again?

From the Google guide · Practice it in a mock interview · Draft my answer from my resume

How would you choose between batch and streaming for a reporting use case?

From the Google guide · Practice it in a mock interview · Draft my answer from my resume

How do you decide what to partition, cluster or index in a large analytical table?

From the Google guide · Practice it in a mock interview · Draft my answer from my resume

How would you make a backfill safe to rerun without double counting?

From the Amazon guide · Practice it in a mock interview · Draft my answer from my resume

When would you choose a columnar format over row storage, and what does it cost you?

From the Amazon guide · Practice it in a mock interview · Draft my answer from my resume

When would you choose streaming over batch for a product metric, and what would it cost?

From the Meta guide · Practice it in a mock interview · Draft my answer from my resume

How would you make a nightly batch job safe to rerun?

From the Microsoft guide · Practice it in a mock interview · Draft my answer from my resume

How would you make a large backfill safe to rerun without double counting?

From the NVIDIA guide · Practice it in a mock interview · Draft my answer from my resume

When would you choose streaming over batch for a product metric, and what does it cost?

From the LinkedIn guide · Practice it in a mock interview · Draft my answer from my resume

How would you make a backfill safe to rerun without double counting?

From the Apple guide · Practice it in a mock interview · Draft my answer from my resume

How would you make a large backfill safe to rerun without double counting?

From the Tesla guide · Practice it in a mock interview · Draft my answer from my resume

Given only the documentation for an unfamiliar API, how would you build a working ingestion job in an hour?

From the Stripe guide · Practice it in a mock interview · Draft my answer from my resume

When would you choose streaming over batch for a financial report, and what does it cost in correctness?

From the Stripe guide · Practice it in a mock interview · Draft my answer from my resume

How would you make a large backfill safe to rerun without double counting?

From the Intel guide · Practice it in a mock interview · Draft my answer from my resume

When would you choose streaming over batch for a monitoring use case, and what does it cost?

From the Intel guide · Practice it in a mock interview · Draft my answer from my resume

This query scans a billion rows and takes twenty minutes. How do you diagnose it?

Practice it in a mock interview · Draft my answer from my resume

Explain the difference between partitioning and clustering, and when each helps.

Practice it in a mock interview · Draft my answer from my resume

A stakeholder wants historical reporting to reflect a customer's current region. What breaks?

Practice it in a mock interview · Draft my answer from my resume

How would you model an event stream that arrives out of order?

Practice it in a mock interview · Draft my answer from my resume

When would you choose a wide table over a normalised model?

Practice it in a mock interview · Draft my answer from my resume

The upstream source sends yesterday's data again. What happens to your job?

Practice it in a mock interview · Draft my answer from my resume

How would you backfill two years of history without disrupting the daily load?

Practice it in a mock interview · Draft my answer from my resume

Late arriving data shows up three days after the partition closed. What do you do?

Practice it in a mock interview · Draft my answer from my resume

What do you monitor, and what pages someone at three in the morning?

Practice it in a mock interview · Draft my answer from my resume

What causes a shuffle, and why is it expensive?

Practice it in a mock interview · Draft my answer from my resume

Your job is slow and one task takes far longer than the rest. What is happening?

Practice it in a mock interview · Draft my answer from my resume

Explain data skew and two ways to handle it.

Practice it in a mock interview · Draft my answer from my resume

When would you choose streaming over a scheduled batch job?

Practice it in a mock interview · Draft my answer from my resume

What does exactly once processing actually guarantee, and where does it not hold?

Practice it in a mock interview · Draft my answer from my resume

How do watermarks handle out of order events in a windowed aggregation?

Practice it in a mock interview · Draft my answer from my resume