General Motors Data Engineer interview typically runs 3 rounds: recruiter email screen, DSA technical round, and system design round. The process moves quickly, sometimes scheduling interviews with only hours of notice.
$102K
Avg. Base Comp
$186K
Avg. Total Comp
3-4
Typical Rounds
1-2 weeks
Process Length
Our candidates report that General Motors — at least through this contract pipeline — cares most about whether you can reason through operational data systems under realistic constraints. The technical content here wasn't abstract: nested key-value storage, transaction-system architecture, complex SQL joins, and MapReduce-style streaming all showed up in a single loop. That breadth suggests the team wants engineers who can move fluidly between data modeling, query optimization, and distributed processing — not specialists who only know one layer.
One thing that stood out from this experience is how much platform familiarity appeared to influence the outcome. The candidate came in with strong Snowflake experience and felt that Databricks exposure may have been the deciding factor. That's a non-obvious signal worth taking seriously. GM's data infrastructure appears to lean toward streaming and processing frameworks, so candidates coming from purely warehouse-centric backgrounds may need to bridge that gap explicitly during the interview — not just list tools, but demonstrate how they'd apply streaming and batch processing thinking to problems like payment pipelines or retailer data warehouses.
The process itself was compressed and, by the candidate's account, inconsistently run — one interviewer seemed underprepared and struggled to clarify the problem statement. We've seen this pattern before in contract-routed roles: the technical bar is real, but the interview experience can vary significantly by interviewer. That means candidates need to be self-directed enough to reframe ambiguous prompts and drive the conversation forward, because you may not always get clean setup from the other side of the call.
Synthesized from 1 candidate report by our editorial team.
Had an interview recently?
Share your experience. Unlock the full guide.
Real interview reports from people who went through the General Motors process.
The process moved pretty fast for me, and that was probably the first thing that stood out. I got a few questions over recruiter email first, then was scheduled for two technical rounds on very short notice — literally about three hours before the interview. It was for a contract role through Optimal Inc., so the whole thing felt a bit more compressed than a typical full-time loop.
The first technical round was data structures and algorithms, and it was more awkward than hard because the interviewer seemed pretty unprepared and couldn’t really explain the problem clearly or answer follow-up questions about it. The main prompt I remember was a nested key-value storing question, which was testing how I’d structure data and handle lookups. The second round shifted into system design and was centered on a transaction system. That one felt more like a practical architecture discussion than a coding exercise. Separately, there was also an SQL exercise in the process that involved writing complex queries with tricky joins, and another streaming-style exercise that focused on data processing and analysis, with some MapReduce-type thinking mixed in. Overall, the technical bar seemed real, but the process itself was pretty inconsistent.
I thought I did well, but I also got the sense that prior Databricks exposure may have mattered more than my background, since most of my experience was with Snowflake. In the end I didn’t get the offer. If you’re preparing, I’d focus on being comfortable with tricky SQL joins and being able to talk through a transaction-system design clearly, since those were the most concrete parts of the loop.
Prep tip from this candidate
Drill complex SQL joins and be ready to design a transaction system at a high level, since those were both explicitly part of the process. It would also help to practice explaining your data platform experience in terms of Databricks versus Snowflake, because that seemed to matter in the evaluation.
Share your own interview experience to unlock all reports, or subscribe for full access.
Sourced from candidate reports and verified by our team.
Topics based on recent interview experiences.
Featured question at General Motors
Design a data warehouse for a new online retailer
| Question | |
|---|---|
| Hurdles In Data Projects | |
| Duplicate Rows | |
| Testing Constraints | |
| Client Solution Pushback | |
| Payment Data Pipeline | |
| Decreasing Tech Debt | |
| Your Strengths and Weaknesses | |
| The Brackets Problem | |
| Google Maps Improvement | |
| Classification and Regression | |
| Target Indices | |
| Transformer Encoder Layer | |
| Losing Users | |
| Ticket Agent Analysis | |
| String Palindromes | |
| Implementing the Fibonacci Sequence in Three Different Methods | |
| International e-Commerce Warehouse | |
| Search Timeout | |
| Matrix Multiplication | |
| Why Do You Want to Work With Us | |
| Cloud-Agnostic Deployments | |
| Analyzing Churn Behavior | |
| Processing Large CSV | |
| LRU Cache 1 | |
| Linear vs Logistic Regression | |
| Empty Neighborhoods | |
| 2nd Highest Salary | |
| Rolling Bank Transactions | |
| Comments Histogram |
Synthesized from candidate reports. Individual experiences may vary.
The process begins with a recruiter reaching out via email with a few preliminary screening questions. Scheduling for technical rounds follows quickly, with very short notice given before interviews — as little as three hours in some cases.
The first technical interview focuses on data structures and algorithms. The main problem involves a nested key-value storage question, testing how candidates structure data and handle lookups efficiently.
The second technical interview is a system design discussion centered on designing a transaction system. It is framed as a practical architecture conversation rather than a live coding exercise.
Candidates also complete an SQL exercise involving complex queries with tricky joins, as well as a streaming-style data processing exercise that incorporates MapReduce-type thinking. These may be administered as separate components within the overall technical loop.