PwC data scientist interviews, as candidates describe them, lean toward business cases and project walkthroughs, with SQL and machine learning questions alongside. Formats vary by report, and total durations of two and four weeks were reported.
$95K
Avg. Base Comp
$110K
Avg. Total Comp
Not reported
Typical Rounds
2-4 weeks
Process Length
Across six candidate reports, the PwC data scientist interview looks closer to a consulting assessment than a pure coding screen. Most candidates describe a business case in some form: analyzing a data acquisition process for an advertising company, solving a real client problem, or presenting a group case to managers. Expect to structure a messy problem and explain your reasoning out loud.
Technical content shows up, but its depth varies. One candidate was asked about SQL and Power BI usage, and another talked through Python, SQL, feature engineering and a RAG project. Another candidate described more theory-heavy rounds: PCA versus LDA, Latent Dirichlet Allocation, debugging Python without running it, and a fraud detection scenario that moved into synthetic data and governance. Another reported overfitting and class imbalance questions, then a slide presentation of a project they were proud of.
Behavioral and manager conversations appear in several accounts, covering strengths and weaknesses, teamwork, conflict, and motivation. Two candidates reported total durations, about two weeks and about four weeks, and another mentioned delays between rounds. Outcomes were mixed: one offer, one declined offer and four no-offer reports. The common thread is being ready to defend a project end to end and explain it to non-technical managers.
Synthesized from 6 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 Pwc process.
What stood out to me most was that the conversation stayed very business-focused rather than turning into a pure coding screen. I was asked to walk through how I usually start by understanding a stakeholder’s problem and translating it into an analytical use case, so I spent a lot of time talking about how I frame business questions before jumping into modeling. From there, the discussion moved into my hands-on work with data preprocessing, feature engineering, and building models in Python and SQL. I also talked through the machine learning methods I’ve used, including regression, classification, and boosting, and then shifted into a few examples from deep learning work for NLP. The interviewer seemed especially interested in how I’ve applied newer GenAI ideas in practice, so I discussed prompt engineering and a RAG-based solution where we integrated LLMs with vector databases to improve response accuracy. I also mentioned collaborating closely with DevOps, since that came up as part of how I’ve worked across teams.
Overall, it felt like they were evaluating whether I could connect technical work to real business impact and communicate that clearly to non-technical stakeholders. It wasn’t framed as a hard algorithmic interview, but it did require being precise about the kinds of problems I’ve solved and the tools I used. I didn’t get an offer, so for me the main takeaway was that this process rewarded candidates who can speak comfortably about the end-to-end workflow, not just modeling in isolation. If I were doing it again, I’d be ready to explain concrete examples of how I’ve taken a business problem from scoping through deployment or handoff, especially where Python, SQL, and GenAI all played a role.
Prep tip from this candidate
Be ready to explain how you turn stakeholder problems into analytical use cases, then tie that directly to examples of preprocessing, feature engineering, and model building in Python and SQL. It also helps to have a concrete RAG or prompt-engineering project ready, since GenAI use cases came up alongside collaboration with DevOps.
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 Pwc
Write a query to return the two students with the closest test scores and the score difference
| Question | |
|---|---|
| Experiment Validity | |
| Size of Joins | |
| Sort Strings | |
| Hurdles In Data Projects | |
| RAG Strict Source Control | |
| Slow SQL Query | |
| Data Pipelines and Aggregation | |
| Data Preparation for Imbalanced Data | |
| RAG Hallucinations | |
| Overfit Avoidance | |
| Electricity Supply | |
| Company Acquisition Choice | |
| Why Do You Want to Work With Us | |
| Your Strengths and Weaknesses | |
| User Journey Analysis | |
| Data Cleaning Experiences | |
| Clustering Basketball Players | |
| Third Party Ad Pricing | |
| Feedback Sentiment Analysis | |
| 2nd Highest Salary | |
| Rolling Bank Transactions | |
| Merge Sorted Lists | |
| Longest Streak Users | |
| Top 3 Users | |
| Raining in Seattle | |
| Maximum Profit | |
| Find the First Non-Repeating Character in a String | |
| Bagging vs Boosting | |
| Revenue Retention |
Synthesized from candidate reports. Individual experiences may vary.
Candidates typically describe an HR call covering background and basic qualifications, or scheduling the next step. One candidate also reports an online assessment before the HR call. Expect to explain your motivation for PwC and your skills briefly.
Several candidates report a business case as the core technical assessment. Examples include analyzing a data acquisition process for an advertising company, proposing a solution to a client problem, and a group case followed by a presentation to managers. Structure your approach and explain trade-offs.
Reports vary. Some candidates describe SQL and Power BI questions, or Python, SQL and GenAI project discussions. Others report ML theory such as overfitting, class imbalance, PCA versus LDA, and Python debugging. Prepare to explain how you used each tool you list on your resume.
One candidate presented slides on a project they were proud of for about 30 minutes, and another was pushed to explain what went wrong in ML projects. It may help to prepare one project with clear trade-offs, failures and recovery steps.
Candidates report a final conversation with a manager, senior manager or partner. Topics included behavioral questions, ambiguity, teamwork, business impact and background. Be ready to link your technical work to business outcomes.
