1 medium Python question
Out loud, in order: clarify, build it, check it, explain it.
Practise PythonLoadingone moment…
Data engineering interview · Big Tech & Enterprise Software
AWS-native, SQL + data-modeling heavy, with Leadership Principles woven through every round and a Bar Raiser.
01
And what its data engineers work with.
02
Stages in the order candidates meet them.
6 stages. 5 rounds have reported questions, described below.
03
The topics it covers and the questions candidates reported.
HackerRank-style
High confidence4 reported questions. See them below
paper, no IDE
Medium confidence2 reported questions. See them below
4 reported questions. See them below
3 reported questions. See them below
3 reported questions. See them below
04
By round. Questions marked Expected are written from the company's standards.
Online AssessmentHigh confidence
Written SQL (variant)Medium confidence
Data ModelingHigh confidence
System DesignHigh confidence
Behavioral (LP / Bar Raiser)High confidence
Nobody reported these exact questions at Amazon. They are the questions a loop for each of its top topics is known to ask, so prepare for them as likely, not as confirmed.
05
Ranked by how much of the evidence points at each topic.
06
Do the first group before the second.
SQL: window functions, ranking, rolling averages, dedup
Do first (3)Data modeling: star/snowflake, SCD types, normal forms, Redshift keys
8–10 reusable STAR stories, each mapped to several Leadership Principles
Spark internals; Python ETL (idempotency, late data)
ThenAWS services (Redshift/S3/EMR/Glue/Kinesis)
Sharding & consistent hashing
07
Why candidates were turned down, as reported.
Weak/under-prepared Leadership Principle stories (the Bar Raiser shares the final hiring decision)
Behavioral is named in 7 of 50 company accounts.Shaky data modeling — wrong SCD or unjustified schema
Modeling is named in 14 of 50 company accounts.Slow SQL under time pressure
SQL is named in 37 of 50 company accounts.08
6 reported. Each opens and closes. Composites are labelled.
This candidate went through six rounds. First an online test with Python, a hard SQL question, 15 multiple-choice SQL questions and leadership questions. Then a written SQL round on a given data model, a round on basics (star against snowflake, Spark, slowly changing dimensions, Kafka, a lending model), system design, a mixed technical and behavioral round, and a final behavioral round.
What to take from it Expect SQL in several forms, and expect leadership questions in every round.
A 2025 post on TeamBlind describes the same shape: a recruiter call, an online test, a loop of interviews and then a Bar Raiser. We only saw the outline of the post, so this is the shape of the loop only.
What to take from it Plan for the Bar Raiser. It is a separate round near the end.
The interviews tested SQL a lot: CTEs, window functions and joins. In one phone screen the SQL was tricky and used window functions, CTEs and subqueries.
What to take from it Practise window functions and CTEs until you can write them without help.
One SQL question gave a sales table with product, year, capital, units and cost per unit. The task was to find the profit for each product in each year.
What to take from it Read the columns first and work out the formula before you write the query.
One candidate had to build a data model from scratch and then answer SQL questions on it. Another said the model design was the hardest technical part.
What to take from it Practise designing a small model out loud, then querying it.
The Bar Raiser round was mostly behavioral. It covered four or five leadership principles, with many follow-up questions. Reports differ on the number of rounds: one has two technical rounds, a hiring manager round and a Bar Raiser, and another has five rounds.
What to take from it Prepare stories for the leadership principles, with numbers, and expect follow-up questions.
High first-hand signal. Leadership Principles are decisive — every technical round also scores LPs, and a Bar Raiser from outside the team shares the final hiring decision.
09
Their reported questions, one at a time, against a stopwatch.
1 medium Python question
Out loud, in order: clarify, build it, check it, explain it.
Practise Python10
Pick your level and the weeks you have.
Built from what Amazon asks, strongest first. For the wider basics at your level, follow the Mid level plan alongside it.
Data modeling
Read: Data modeling chaptersDrill: Modeling practiceSay it aloud: Data modeling questionsFull plan: Ace the modeling round
Behavioral and communication
Read: Behavioral storytellingSay it aloud: Behavioral questions
Python and coding
Read: Python chaptersDrill: Python practiceSay it aloud: Python questionsFull plan: Ace the Python round
Cloud and tools
Rehearse the Amazon loop
Run the question deckThen repeat your weakest topic once more.