>
Uber
Uber Data Scientist Interview Guide 2026
Complete Uber Data Scientist interview guide. Learn about the interview process, key topics in experimentation, SQL, ML, and marketplace analytics for Uber's data science roles.
6 min read
Updated Jun 2026
232+ practice questions
232+
Practice Questions6
Rounds7
Categories6 min
ReadTL;DR
Uber's Data Scientist interview in 2026 reflects the company's data-rich, marketplace-driven business. The process typically includes a recruiter screen, a technical phone screen, and a virtual onsite with four to five rounds. What makes Uber's DS interview unique is the focus on marketplace dynamics, experimentation at scale, and causal inference. You'll need strong SQL skills, solid statistics knowledge, and the ability to think about two-sided marketplace problems. Uber relies heavily on experimentation to make product decisions, so understanding A/B testing, switchback experiments, and interference effects is essential. The product analytics round often involves diagnosing metric movements in a marketplace context, which requires different intuition than single-sided products.
4-6 weeks
232+ questions
Sample Questions
232+ in practice bank
Write a SQL query to calculate driver utilization rate by city for the last quarter
Given tables for trips, drivers, and cities, calculate the percentage of time drivers spend actively on trips versus available but idle, broken down by city.
Write a SQL query to find the top 10 drivers with the highest ratings who have completed at least 100 trips
Join trip and rating tables, filter by trip count, and rank by average rating. Handle edge cases like drivers with only recent trips versus long tenure.
How would you design an experiment to test a new surge pricing algorithm?
Discuss why standard A/B testing fails for marketplace experiments. Explain switchback design, interference between treatment and control, and how you'd measure the effect on both riders and drivers.
Explain the difference between a switchback experiment and a standard A/B test
Discuss why marketplace experiments require special designs. Explain how interference between riders and drivers invalidates standard randomization, and how switchback designs address this.
How would you measure the impact of a new driver incentive program?
Discuss causal inference challenges: selection bias, spillover effects, and long-term vs. short-term impacts. Propose an experimental or quasi-experimental approach to isolate the incentive's effect.
Explain how you'd handle multiple comparisons in an experiment with 10 variants
Discuss Bonferroni correction, false discovery rate control, and when to use each. Explain the tradeoff between statistical rigor and the ability to detect real effects.
Rider cancellation rate increased by 8% last week. How would you investigate?
Walk through a systematic debugging approach. Segment by city, time of day, ride type, driver characteristics, and app version. Consider marketplace effects like supply-demand imbalance.
What metrics would you use to evaluate Uber Eats marketplace health?
Define metrics for all three sides of the marketplace: eaters, restaurants, and delivery partners. Discuss leading vs. lagging indicators and how they interact.
How would you build a model to predict ride demand in a city?
Discuss feature engineering (time, weather, events, historical patterns), model selection, and how you'd evaluate predictions. Explain how you'd handle spatial and temporal granularity.
Tell me about a time you had to communicate a counterintuitive finding to stakeholders
Share a specific example where your data analysis contradicted expectations. Explain how you validated the finding and persuaded decision-makers to act on it.
About the Interview Process
Uber's Data Scientist interview is tailored to their marketplace business. The company needs data scientists who can navigate the complexities of two-sided markets, design experiments that account for interference effects, and translate data into product decisions. Strong SQL skills are the baseline; marketplace intuition is the differentiator.
Recruiter Screen
Initial conversation about your background, the specific team, and the role. Be ready to discuss your experience with marketplace analytics, experimentation, or causal inference.
Technical Phone Screen
SQL problems and basic statistics questions. Expect one to two SQL problems of medium difficulty, possibly followed by a probability or hypothesis testing question.
Onsite: SQL & Data Manipulation
Advanced SQL problems involving window functions, CTEs, self-joins, and complex aggregations. Some teams also test Python/Pandas skills. The queries often reflect real Uber data scenarios.
Onsite: Product Analytics
Given a marketplace scenario, define metrics, diagnose problems, or evaluate a product change. They want structured reasoning that accounts for both supply-side and demand-side effects.
Onsite: Experimentation & Statistics
Deep dive into experimental design, causal inference, and statistical methods. Expect questions about switchback experiments, difference-in-differences, and interference in marketplace settings.
Onsite: Behavioral
Structured behavioral interview focusing on impact, collaboration, and communication. Uber wants data scientists who drive decisions, not just deliver reports.
Timeline
4 to 6 weeks from first recruiter contact to offer.
Tips
Study marketplace experimentation. Standard A/B testing intuition breaks down in two-sided markets.
Practice SQL with window functions and complex joins. Uber's SQL questions are practical and challenging.
Understand the dynamics of supply and demand in a marketplace. Many product analytics questions hinge on this.
Prepare examples of analyses where you influenced product or business decisions with measurable outcomes.
Learn about switchback experiments and difference-in-differences. These come up frequently in Uber DS interviews.
What they test
Uber's Data Scientist interview evaluates four core competencies: SQL fluency, statistical rigor, marketplace intuition, and communication.
SQL is the foundation. You'll face practical queries involving trips, drivers, riders, and marketplace metrics. Expect window functions, CTEs, and multi-table joins that mirror real production queries.
The experimentation round goes deeper than most companies. Uber operates a two-sided marketplace where standard A/B testing breaks down due to interference effects. You need to understand switchback experiments, cluster randomization, and causal inference methods like difference-in-differences.
Product analytics at Uber requires thinking about both sides of the marketplace simultaneously. When a metric moves on the rider side, it often has supply-side causes or consequences. Interviewers test whether you can reason about these interactions.
Marketplace experimentation at Uber
Understanding marketplace experimentation is critical for Uber DS interviews. In a standard product A/B test, you randomly assign users to treatment or control. But in a marketplace, treating some riders differently affects drivers too, creating interference between groups.
Uber addresses this with switchback experiments: instead of randomizing by user, they randomize by time and geography. For example, a city might see the new pricing algorithm for one hour, then the old one for the next hour. This controls for interference but introduces new challenges around carry-over effects and reduced statistical power.
You should also understand difference-in-differences for quasi-experimental settings, and regression discontinuity for policy changes. These methods come up when randomized experiments are impractical, which happens more often than you'd think in a marketplace.
Leveling & Compensation
| Level | Title | YoE | Total Comp (USD/yr) |
|---|---|---|---|
DS3 | Data Scientist I | 0-2 yrs | $130k - $220k |
DS4 | Data Scientist II | 2-5 yrs | $200k - $340k |
DS5 | Senior Data Scientist | 5-10 yrs | $290k - $500k |
DS6 | Staff Data Scientist | 8-15 yrs | $400k - $680k |
Data Scientist I
Strong SQL and statistics fundamentals. Can run analyses and support experiments independently. Delivers clear, actionable insights to the team.
Data Scientist II
Owns the analytics for a product area. Designs and runs experiments end to end. Proactively identifies opportunities and risks in the data.
Senior Data Scientist
Shapes the analytics strategy for a product area. Develops novel methodologies for marketplace experimentation. Mentors junior data scientists and influences cross-functional roadmaps.
Staff Data Scientist
Sets the data science direction for a major product area. Drives company-wide best practices for experimentation and causal inference. Recognized as a thought leader.
How to Stand Out
Behavioral Focus Areas
Impact through data: demonstrating how your analysis directly influenced product or business decisions
Marketplace thinking: showing you understand the dynamics of two-sided platforms
Rigor: maintaining statistical discipline even when stakeholders want quick answers
Communication: translating complex findings into actionable recommendations for product teams
Ownership: proactively identifying problems and opportunities rather than waiting for requests
1.
Uber's marketplace creates unique analytical challenges. Study two-sided market dynamics before your interview.
2.
Practice SQL daily with real-world scenarios. Window functions, CTEs, and date manipulation are essential.
3.
Learn about switchback experiments, CUPED variance reduction, and interference testing. These are Uber staples.
4.
For product analytics, always consider both the rider and driver perspective. One-sided analysis won't score well.
5.
Quantify your impact in behavioral stories. Uber cares about measurable outcomes, not just interesting analyses.
6.
If you have experience with geospatial data analysis, highlight it. Many Uber DS problems involve location data.
7.
Practice explaining causal inference methods to a non-technical audience. Clarity is as important as correctness.
Recommended Resources
Causal Inference: The Mixtape by Scott Cunningham
Trustworthy Online Controlled Experiments by Ron Kohavi
Uber Engineering Blog
FAQ
What makes Uber's DS interview different from Meta or Google?
The biggest difference is the marketplace focus. Uber's product analytics and experimentation questions require you to think about two-sided market dynamics, which most other tech companies don't test. Standard A/B testing knowledge is necessary but not sufficient. You need to understand interference, switchback designs, and supply-demand interactions.
Do I need to know machine learning for Uber's DS role?
It depends on the team. Core marketplace analytics roles focus more on experimentation, causal inference, and product analytics. ML-heavy roles (like demand forecasting or pricing) expect deeper ML knowledge. For most DS positions, understanding regression, classification basics, and feature engineering is enough.
What SQL topics should I focus on?
Window functions (ROW_NUMBER, RANK, LAG, LEAD, running aggregates), CTEs, self-joins, date and time manipulation, and complex GROUP BY queries. Uber's SQL questions are practical and often involve trip-level, driver-level, or marketplace-level metrics.
How important is knowledge of causal inference?
Very important for senior roles. Uber's marketplace creates situations where observational data is misleading due to confounding. Understanding difference-in-differences, instrumental variables, regression discontinuity, and propensity score matching will set you apart from other candidates.
What's the salary range for Data Scientists at Uber?
Total compensation ranges from roughly $130K to $220K at DS3 (entry level), $200K to $340K at DS4, $290K to $500K at DS5 (senior), and $400K to $680K+ at DS6 (staff). These include base salary, RSUs, and bonus. Location and team can affect the range.
Should I prepare differently for Uber Eats versus Rides DS roles?
The core skills (SQL, statistics, experimentation) are the same, but the domain context differs. Uber Eats is a three-sided marketplace (eaters, restaurants, couriers), which adds complexity. Rides focuses on real-time matching, pricing, and ETAs. Tailor your preparation based on which team you're interviewing with.
Comments
