Candidates describe Revolut's Data Scientist interview as a multi-stage process: an HR screen, pandas and SQL coding, ML and statistics theory, then case, bar raiser or manager conversations. Reports of the number of rounds vary.
$124K
Avg. Base Comp
$161K
Avg. Total Comp
3-6 rounds
Typical Rounds
Not reported
Process Length
Candidates describe the Revolut Data Scientist process as technical early and heavier on pandas, SQL and ML fundamentals than a typical analytics interview. The usual opening is a recruiter or HR conversation, and some candidates also report an online assessment with SQL, statistics or aptitude questions. Live coding follows. Several candidates describe working in pandas on a dataset, often with an open-ended business decision, where interviewers want you to ask questions, state assumptions, justify choices aloud and handle edge cases. One candidate was asked whether a solution would still work on a dataset ten times bigger. Another reported being allowed to use AI tools. SQL tasks reported include joins, window functions and transaction-style tables with users, payments and spending trends.
The theory round is where some candidates struggled. Two candidates describe math-heavy questions covering Bayes-style probability, maximum likelihood for Bernoulli or binomial distributions, A/B-test sample sizing and linear versus logistic regression. One also reports questions on the Gauss-Markov theorem, gradient boosting and SGD. Other reports mention a machine-learning system design question, a product case about a metric drop, and a deep dive into past projects. Later stages may include a bar raiser or hiring manager conversation. Reported round counts range from three to six, so ask your recruiter what your own loop contains. Outcomes in the reports are mixed, and some candidates describe slow or missing communication.
Synthesized from 25 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 Revolut process.
The candidate reports six rounds. The first was an HR round on the resume, compensation and notice period. The next was a Python and pandas round, followed by a round on mathematical theorems. After that came a system design round, a bar raiser round and a hiring manager round. The candidate was still studying for the second and third rounds. HR feedback in the first round flagged microphone and network quality as a negative, and the candidate was warned it is essential to present yourself well.
Questions asked: The first technical round was a dataset exercise with AI: clean data in pandas and walk through it. Tasks included joining tables in pandas, finding the most frequent recurring transactions, and calculating retention, app usage and other product metrics to analyze the product.
Prep tip from this candidate
Check your microphone and network connection before the first call, and practice pandas joins and product metrics such as retention.
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 Revolut
Write a query that returns all neighborhoods that have 0 users.
| Question | |
|---|---|
| Rolling Bank Transactions | |
| Top Three Salaries | |
| Experiment Validity | |
| Last Transaction | |
| Like Tracker | |
| Month Over Month | |
| Button AB Test | |
| Daily Logins | |
| Top 3 Users | |
| Third Purchase | |
| Retailer Data Warehouse | |
| Rolling Average Steps | |
| Total Spent on Products | |
| Find the First Non-Repeating Character in a String | |
| Daily Retention Summary | |
| Size of Joins | |
| Post Composer Drop | |
| Cumulative Reset | |
| Google Maps Improvement | |
| Declining Applicants | |
| Suspicious Card Activity | |
| Payments Received | |
| Subscription Retention | |
| Time on FB Distribution | |
| Sort Strings | |
| Consecutive Spending Growth | |
| Hurdles In Data Projects | |
| Success Measurement | |
| Word Frequency |
Synthesized from candidate reports. Individual experiences may vary.
Candidates report a first conversation on their resume, projects, tech stack, motivation for Revolut, salary expectations and notice period. Some describe light technical questions, such as pandas and SQL join choices or how they would take a model to deployment. Practice a concise project walkthrough and a specific answer for why Revolut. One candidate was also told to present themselves well, including audio and network quality.
Some candidates report an early test rather than a call. Reported formats include a HackerRank with two SQL questions and several statistics and A/B-testing questions, and a screen with Python, SQL and multiple-choice ML and statistics questions. Another candidate describes an aptitude test. This step may not apply to every candidate, so confirm the format with your recruiter.
Several candidates describe a roughly one-hour live session with a dataset, often in pandas, and sometimes SQL on tables like users and transactions. Tasks reported include joins, groupby, window functions, cleaning data, labeling data and designing a metric. Interviewers typically want you to ask clarifying questions, justify choices aloud and reason about edge cases. Some candidates report that AI tools were allowed, while another reports LeetCode-style problems without them.
Candidates report a technical round on machine learning fundamentals, with questions that may require deriving or working through the math. Topics reported include Bayes-style probability, Bernoulli or binomial maximum likelihood, A/B-test sample size, linear and logistic regression, loss functions and boosting. Some report a specialization-dependent deep dive or an ML system design question. Review the underlying math as well as the concepts.
Later stages vary between reports. Candidates mention a product or root-cause case, such as diagnosing a metric drop, a deep dive into past projects, a character or fitness conversation, a bar raiser and a hiring manager round. Prepare to break ambiguous problems into hypotheses and data requests, and to explain your own impact on past projects. Reported timelines and communication differ, so ask for your loop in writing.
