Three candidates describe ZS data science interviews built around a recruiter screen, a case study or live data task, and questions about how they approached the analysis. All three reported negative outcomes.
$135K
Avg. Base Comp
$175K
Avg. Total Comp
Not reported
Typical Rounds
Not reported
Process Length
The three candidates who describe ZS data science interviews all faced a case study or data task, and each had to explain how they approached it. One interviewed for a Data Scientist role, one for an intern role and one for an associate role, so the details are related but not identical.
The Data Scientist candidate describes a 30-minute phone screen, an online assessment covering Python, SQL, ML theory and probability/stats, an unstructured case study with SQL questions, and a final round with a one-hour case study plus 45 minutes of behavioral questions. They found the case conceptual and reliant on communication and framing. They also struggled to follow one interviewer and asked many clarifying questions, which they suspect hurt them.
The intern candidate completed a 2-hour independent case after a short business overview, then spent 45 minutes walking an interviewer through their code. Questions focused on why they chose certain algorithms and how those algorithms work internally. The associate candidate worked through a live data science task, did not pass it, and was asked why they had not tried a different technique to get closer to a complete model.
All three reported negative outcomes, so preparation should center on the gaps they describe. Have your coding environment working before you start, and be ready to explain every modeling decision and the alternatives you considered. Practice SQL, know your past projects well, and confirm your reading of a prompt early if it is unclear. One candidate also noted that ZS works across time zones, so double-check meeting times.
Synthesized from 3 candidate reports by our editorial team.
Had an interview recently?
Share your experience. Unlock the full guide.
Real interview reports from people who went through the Zs Associates process.
I applied with two referrals, which I thought gave me a decent shot. The process started with a quick 30-minute phone screening to assess my background and interest. It was pretty standard—mostly conversational, no coding involved.
The second round was an online assessment covering Python, SQL, ML theory, and probability/stats. The questions tested practical coding ability and SQL proficiency, plus conceptual knowledge like how to handle outliers in datasets and understanding different ML algorithms and their evaluation metrics. This was more technical and time-pressured than the initial call.
Third round shifted to an unstructured case study paired with more SQL questions. I had to work through a business problem, think on my feet, and demonstrate how I'd tackle it with code and data logic. The case study itself wasn't overly difficult—it was pretty conceptual and relied more on communication and framing than heavy prep. The SQL component was straightforward too.
The final round was structured into two parts: a one-hour case study and then 45 minutes of behavioral questions. The case study here was more formal than round three, and we went deeper into the analysis and decision-making. The behavioral portion covered my past projects, relevant experience, and how I'd approached technical challenges. One thing I underestimated was how much the interviewer and communication style mattered. I'm a native English speaker, but I had a hard time understanding my case study interviewer in the final round—their accent and pacing threw me off. I found myself asking a lot of clarifying questions, which in hindsight might have cost me. It's a reminder that even when you know the material, the interview dynamic and your ability to understand and connect with the interviewer is crucial.
Prep tip from this candidate
Go deep on your past projects and be ready to explain them clearly—interviewers will probe your actual experience. Practice SQL window functions and time-series handling, since those came up consistently. For case studies, focus on clear communication and understanding your interviewer; if something doesn't land, confirm your interpretation early rather than making assumptions.
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 Zs Associates
What are the assumptions of linear regression?
| Question | |
|---|---|
| Late Deliveries | |
| Data Preparation for Imbalanced Data | |
| Your Strengths and Weaknesses | |
| Docs Metrics | |
| PCA and K-Means | |
| Interpreting Fraud Detection Trends | |
| 2nd Highest Salary | |
| Top Three Salaries | |
| Longest Streak Users | |
| Rolling Bank Transactions | |
| Closest SAT Scores | |
| Employee Salaries | |
| Largest Salary by Department | |
| Merge Sorted Lists | |
| Top 3 Users | |
| Top 5 Turnover Risk | |
| First to Six | |
| Size of Joins | |
| Prime to N | |
| Find the Missing Number | |
| Fair Coin | |
| Duplicate Rows | |
| Experiment Validity | |
| Paired Products | |
| Get Top N Frequent Words | |
| First Touch Attribution | |
| Manager Team Sizes | |
| SELECTive Wine Connoisseur | |
| Bagging vs Boosting |
Synthesized from candidate reports. Individual experiences may vary.
Candidates report a short opening call with a recruiter. One describes about 30 minutes of conversation with no coding, another recalls behavioral questions about their background and interest in consulting, and a third was asked about availability and relocation. Expect a conversational screen rather than a technical test.
One candidate describes an online assessment covering Python, SQL, ML theory and probability/stats, including outlier handling and evaluation metrics. Another reports a fit interview to gauge cultural alignment. This stage differed between candidates and roles, so confirm the format with your recruiter.
Candidates report a case study or hands-on data task. Formats included an unstructured case paired with SQL questions, a 2-hour independent case after a 10-minute overview, and a live task done during the interview. Expect to work through data, make modeling choices and explain your reasoning as you go.
The intern candidate describes a 45-minute session walking through their code. The interviewer probed why they chose their algorithms, how those work internally and what assumptions they made. Another candidate was asked in a debrief why they had not tried a different technique. Be prepared to defend trade-offs and alternatives.
Two candidates describe a final stage with behavioral questions: one had 45 minutes after a one-hour case study, and the intern candidate had a 30-minute interview on why they wanted consulting. Topics included past projects, relevant experience and technical challenges. The intern's final round also returned to questions about the earlier case.
