Multi-join queries, window functions, CTEs
Out loud, in order: clarify, build it, check it, explain it.
Practise SQLLoadingone moment…
Data engineering interview · Finance, Payments & Fintech
CoderPad SQL + Python, distributed-systems fundamentals (RDD vs DF, why PySpark), data modeling, and a hiring committee.
01
And what its data engineers work with.
02
Stages in the order candidates meet them.
6 stages. 3 rounds have reported questions, described below.
03
The topics it covers and the questions candidates reported.
medium–hard
High confidence3 reported questions. See them below
2 reported questions. See them below
3 reported questions. See them below
04
By round. Questions marked Expected are written from the company's standards.
SQLHigh confidence
Python / CoderPadMedium confidence
Distributed systems / SparkMedium confidence
Nobody reported these exact questions at Goldman Sachs. 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 medium–hard (multi-join, window, CTE, optimization)
Do first (3)Python CoderPad (arrays/trees/recursion/graphs)
Spark/distributed fundamentals (RDD vs DF, HDFS/YARN, why PySpark, Java vs Scala)
Data modeling + ETL architecture
ThenBehavioral/values
07
Why candidates were turned down, as reported.
Weak SQL under medium–hard load
SQL is named in 37 of 50 company accounts.Shaky Spark/distributed fundamentals (RDD vs DF, why PySpark)
Spark is named in 25 of 50 company accounts.Shallow data modeling
Modeling is named in 14 of 50 company accounts.08
3 reported, 2 composite. Each opens and closes. Composites are labelled.
A recruiter call, a CoderPad round with SQL and Python, a round on distributed systems (Spark, HDFS, YARN, RDD against DataFrame), a round on data modeling and architecture, then behavioral.
What to take from it Know how Spark runs on a cluster, not only its API.
A HackerRank round with stack and sliding-window problems, a CoderPad round on trees and a dynamic-programming grid problem, then a super day with heap, queue and sliding-window problems and a talk about your current projects.
What to take from it Goldman can test data structures hard. Practise sliding window, trees and heaps.
A 2020 thread tells candidates to expect data structures and algorithms questions, so the coding bar can be as high as for software roles.
What to take from it Prepare algorithms as well as SQL.
Composite walkthroughs. Built from the company's documented loop, not from one person's account.
This is a composite, not one person's account. It is built from the loop we have on record for Goldman Sachs. medium–hard. Questions reported for this round: Multi-join queries, window functions, CTEs; Performance optimization; Top-N, running totals, gaps-and-islands.
What to take from it Practise these questions before you reach this round.
This is a composite, not one person's account. It is built from the loop we have on record for Goldman Sachs. Questions reported for this round: Data manipulation + algorithms; Arrays, trees, recursion, simple graph traversal.
What to take from it Practise these questions before you reach this round.
Strong SQL + Spark fundamentals + data modeling; explain logic and adapt as requirements change.
09
Their reported questions, one at a time, against a stopwatch.
Multi-join queries, window functions, CTEs
Out loud, in order: clarify, build it, check it, explain it.
Practise SQL10
Pick your level and the weeks you have.
Built from what Goldman Sachs asks, strongest first. For the wider basics at your level, follow the Mid level plan alongside it.
Spark and big data
Read: Spark chaptersDrill: Spark playgroundSay it aloud: Spark questionsFull plan: Ace the Spark round
Python and coding
Read: Python chaptersDrill: Python practiceSay it aloud: Python questionsFull plan: Ace the Python round
Behavioral and communication
Read: Behavioral storytellingSay it aloud: Behavioral questions
Rehearse the Goldman Sachs loop
Run the question deckThen repeat your weakest topic once more.