Revolut Data Scientist interview questions & guide 2026

At Revolut, data science is not an isolated research function—it is the operational core of a global financial super-app serving tens of millions of users. Revolut operates with extreme pace, high technical rigor, and a deeply quantitative, metrics-driven engineering culture. A Data Scientist here builds production-grade statistical and machine learning systems that power core product features, including real-time fraud detection algorithms, automated credit decisioning models, algorithmic trading flows, dynamic FX risk pricing, and hyper-personalized user retention campaigns.

Because Revolut scales across diverse regulatory jurisdictions and processes billions of daily transactions, the data challenges are vast. You will work on massive-scale transaction streaming, user behavior analytics, and complex product mechanics where small statistical improvements unlock massive financial outcomes. The organization values extreme autonomy, mathematical precision, and rapid execution. Candidates who thrive in this role combine strong theoretical foundations in probability and machine learning with pristine SQL and algorithm skills, alongside sharp product intuition.

This interview process is rigorous, rapid, and heavily technical. It tests your ability to translate abstract mathematical principles—such as Maximum Likelihood Estimation and Bayesian inference—into scalable production code, while maintaining a clear eye on business impact and product metrics.

Expect a demanding combination of live data manipulation, mathematical derivations, statistical modeling, and product case studies. Questions are drawn directly from candidate experiences across Revolut's global loops and reflect the high bar set for quantitative rigor and practical execution.

SQL & Data Manipulation

This category tests your ability to query large-scale transactional datasets efficiently under real-time constraints, perform complex aggregations, and compute dynamic metrics using core SQL features.

• Write a query using SQL window functions (SUM() OVER, DENSE_RANK()) to calculate a rolling 7-day transaction volume per user and rank top spenders within each country.
• Given a table of card transactions and risk flags, identify all user accounts where transaction velocity spiked by more than 300% week-over-week.

Preparation for Revolut requires balancing academic mathematical rigor with fast execution in SQL and Python. You must demonstrate strong fundamental theory—proving that you truly understand the algorithms you use rather than treating them as black boxes—while proving that you can write clean, production-ready code.

Role-related knowledge – You must know classical statistics, machine learning theory, Python, and SQL cold. Expect to derive core loss functions (e.g., MLE, cross-entropy), solve probability puzzles on paper, write complex SQL aggregations under time pressure, and write idiomatic Python code.

Problem-solving ability – Interviewers evaluate how structured and methodical you are when faced with ambiguous product problems or system architectures. You should articulate clear hypotheses, break down complex ML design problems logically, state assumptions explicitly, and systematically walk through edge cases.

Execution pace & communication – Revolut operates with high velocity. Interviewers often use a rapid-fire questioning style. Demonstrate conciseness, technical clarity, and focus. Avoid long-winded answers, stay responsive to hints, and pivot quickly when redirected.

Culture fit & ownership – Revolut prioritizes individual ownership, high performance, and direct communication. Show that you take responsibility for end-to-end execution, are comfortable operating without hand-holding, and thrive in metric-driven, fast-changing environments.

The Data Scientist hiring process at Revolut is fast-paced, multi-staged, and rigorous. The loop is heavily weighted toward live technical evaluation, mathematical theory, and live coding. While recruiters maintain transparent communication throughout the pipeline, candidates should prepare for demanding technical rounds that leave little room for vagueness.

You will encounter clear progression gates: an initial recruiter screen, followed by live coding and SQL evaluations, deep-dive classical machine learning theory rounds, an ML system design or case study round, and a final seniority and cultural assessment (often referred to as a bar raiser).

This visual timeline outlines the typical sequence from screening to final offer. Use this roadmap to structure your preparation systematically—focusing first on core algorithms, SQL, and mathematical derivations before shifting energy toward system design and behavioral alignment.

SQL & Data Manipulation

The live SQL evaluation at Revolut goes far beyond basic queries. You must demonstrate the ability to manipulate granular financial event logs, execute windowed aggregations, and compute complex business metrics efficiently under time constraints.

Be ready to go over:

• SQL window functions – Master positional functions (LEAD, LAG), ranking functions (ROW_NUMBER, DENSE_RANK), and dynamic aggregate framing (OVER (PARTITION BY ... ORDER BY ... ROWS BETWEEN)).
• Data aggregation & metrics – Combining CASE WHEN logic inside aggregation functions (COUNT(DISTINCT ...), SUM()) to calculate complex conditional conversion rates and metric ratios.

As a Data Scientist at Revolut, you sit at the intersection of quantitative modeling, software engineering, and product strategy. You are directly responsible for designing, deploying, and maintaining models and analytical systems that impact millions of users across the global financial app.

• Deploy Production Machine Learning Models: Design, train, evaluate, and deploy real-time ML systems—ranging from XGBoost ensemble models to deep learning architectures—powering credit scoring, fraud prevention, dynamic pricing, and user personalization.
• Drive Product Analytics & Experimentation: Partner closely with Product Managers and Software Engineers to design rigorous A/B experiments, establish key operational metrics, and continuously optimize core conversion funnels.
• Build & Maintain Scalable Data Pipelines: Author pristine, performant SQL queries and Python data pipelines to aggregate massive streams of financial data, ensuring feature stores remain synchronized across batch and real-time inference engines.
• Perform Rigorous Root-Cause Analyses: Troubleshoot sudden operational shifts—such as sudden changes in approval rates, unexpected metric drops, or algorithmic anomaly spikes—by querying granular logs and applying diagnostic statistical techniques.
• Communicate Insights & Influence Strategy: Translate complex statistical trade-offs, model risk profiles, and experimental conclusions into actionable strategies for cross-functional product leads and executive leadership.

Candidates must possess strong technical foundations, fast execution skills, and comfortable working with complex mathematical concepts.

Technical Skills

• Core Programming: Fluent in Python (Pandas, NumPy, Scikit-learn, PyTorch/TensorFlow) with strong algorithmic problem-solving skills (data structures, recursion, dynamic programming).
• Data Manipulation: Exceptional SQL skills, including advanced window functions, conditional aggregations, dynamic CTEs, and query optimization on modern warehouse engines.
• Mathematics & Statistics: Mastery of probability theory, Bayesian inference, Maximum Likelihood Estimation (MLE), linear algebra, and hypothesis testing.
• Machine Learning: Deep theoretical and practical knowledge of supervised/unsupervised learning, tree ensembles (XGBoost, LightGBM), regularization (L1/L2), model evaluation metrics, and ML system design.

Experience & Qualifications

• Must-have skills:

  • Proven track record (typically 2+ years for core roles, 6+ years for senior roles) building and shipping production machine learning or statistical systems.
  • Demonstrated experience writing clean Python code and advanced SQL queries under live coding conditions.
  • Strong theoretical foundation in classical probability, mathematical statistics, and ML model mechanics.
  • Direct experience designing, launching, and analyzing online A/B experiments.

• Nice-to-have skills:

  • Prior experience in fintech, banking, card payments, fraud detection, or credit risk domains.
  • Hands-on experience with real-time streaming architectures (Kafka, Flink) and enterprise feature stores.
  • Experience in specialized DS domains such as NLP (for automated customer support/text analytics) or computer vision (for identity verification/KYC).

Q: How difficult are the live coding and math rounds compared to other tech companies? The technical threshold at Revolut is high. Unlike companies that evaluate high-level framework usage, Revolut tests fundamental mathematics—requiring live proofs and manual derivations of probability problems and MLE loss functions—alongside time-constrained Python algorithm challenges.

Q: Does Revolut allow the use of AI coding assistants during live coding screens? No. Live coding interviews are conducted in standardized environments where AI coding assistants or autocompletion tools are prohibited. You are expected to write syntactically sound Python and SQL manually while communicating your thought process clearly.

Q: How quickly does the interview loop progress? The hiring loop moves fast. Recruiters generally provide updates within a few business days following each stage, and the entire multi-stage process is often completed within 3 to 4 weeks depending on scheduling availability.

Q: What is the primary work culture attribute Revolut evaluates for? Revolut heavily values high ownership, extreme speed, logical rigour, and a bias for action. Candidates who succeed demonstrate direct communication, high resilience under time pressure, and a passion for driving measurable metrics rather than passive theoretical work.

• Practice Pen-and-Paper Math Derivations: Do not rely on high-level libraries during your prep. Practice deriving MLE for Binomial/Poisson distributions, computing Bayes' theorem problems, and writing out gradient updates for cross-entropy loss by hand or on a digital whiteboard.
• Master SQL Window Functions Under Time Pressure: Ensure you can write advanced SQL queries fluently. Focus on multi-level partitioning, frame specs (ROWS BETWEEN), and conditional aggregations without hesitation.

• Be Concise and Direct: Revolut interviewers appreciate punchy, structured answers. When asked a theoretical or product question, state your answer directly, back it up with mathematical or empirical logic, and avoid long introductions.
• Structure Your ML System Design: Use a clear framework for ML design rounds: Clarify Requirements $\rightarrow$ Frame as ML Problem $\rightarrow$ Data Ingestion & Features $\rightarrow$ Model Selection & Training $\rightarrow$ Offline/Online Evaluation $\rightarrow$ Deployment, Latency & Monitoring.

A Data Scientist role at Revolut offers an exciting opportunity to build high-impact financial products at global scale. You will solve complex quantitative problems in real-time fraud detection, credit underwriting, trading automation, and user engagement across tens of millions of active users.

To succeed in this interview loop, focus your preparation on mathematical foundations, fast live coding in Python, pristine SQL execution, and structured ML system architecture. Approach your prep methodically: review theoretical ML proofs, drill LeetCode Medium algorithms, practice complex windowed SQL queries, and hone your metric design skills.

You can explore additional interview insights, practice questions, and detailed preparation resources on Dataford to refine your technical preparation and boost your confidence.

This compensation module outlines representative salary ranges and equity structures for data science roles at Revolut. Base pay, performance bonuses, and equity grants scale significantly with seniority, geographic location, and track record. Use this data to set clear expectations during initial recruiter discussions around compensation.