INTERVIEW GUIDE

Uber Data Scientist Interview: Questions & Process

Uber's data scientist interview is product- and experimentation-focused, reflecting its two-sided marketplace. Expect a recruiter screen, a technical screen (SQL plus stats/probability), and a loop covering analytical SQL, A/B testing and experiment design, a product-metrics case, and behavioral. They want people who can define the right metrics, design and read experiments, and turn marketplace data into decisions.

The interview process

1. Recruiter screen — ~30 min call

Tests: background, why Uber, role fit, logistics

2. Technical phone screen — 45-60 min

Tests: SQL queries plus probability/statistics fundamentals

3. Analytics / SQL round — 45 min

Tests: complex SQL, metric computation, and interpreting results

4. Experimentation / product-metrics round — 45 min

Tests: A/B test design, statistical reasoning, and defining metrics for a feature

5. Behavioral / hiring manager — 45 min

Tests: communication, collaboration with product/eng, and how you drive impact with data

Questions you're likely to get

Technical

Write a query to compute day-1 driver retention from a trips table.
Using a window function, find each city's top three drivers by completed trips.
Calculate the share of orders delivered on the same day they were requested.
Find riders whose temperature of activity rose week over week from an events table.
Given two arrays of trip values, return indices of two that sum to a target.
Explain p-values and statistical significance to a non-technical product manager.

Role-specific

Design an A/B test for a new pricing feature — what's your metric, sample size, and guardrails?
Rider conversion dropped 3% in one city last week — how do you investigate?
What metrics define a healthy two-sided marketplace, and how do they trade off?
How would you measure the impact of surge pricing on driver supply and rider demand?
An experiment shows a tiny but significant lift — would you ship it? Why or why not?

Behavioral

Tell me about a time your analysis changed a product decision.
Describe a time you had to explain a statistical result to a skeptical stakeholder.
Tell me about a time you worked with engineers and PMs to ship something.

Practice these problems live

Relevant LeetCode problems for the Uber Data Scientist loop. Start a live, AI-run coding interview on any of them — or open the problem on LeetCode.

▶ Game Play Analysis IV· Medium · LeetCode #550

Game play analysis IV maps directly to the retention metrics Uber data scientists compute.

▶ Immediate Food Delivery II· Medium · LeetCode #1174

Immediate food delivery II tests conditional aggregation like same-day delivery rates.

▶ Department Top Three Salaries· Hard · LeetCode #185

Department top three salaries practices the window/rank pattern for top-N-per-group queries.

▶ Rising Temperature· Easy · LeetCode #197

Rising temperature tests date-based self-joins common in event analysis.

▶ Two Sum· Easy · LeetCode #1

Two Sum covers the light DSA warm-up that can appear alongside SQL.

▶ Maximum Subarray· Medium · LeetCode #53

Maximum subarray is a simple algorithm check for the occasional coding portion.

How to answer (worked examples)

Design an A/B test for a new pricing feature.

Structure it: state the hypothesis and a single primary metric, then randomization unit (rider, driver, or city — watch for marketplace interference), sample size and duration from a power calculation, and guardrail metrics so you don't help one side while hurting the other. Discuss network effects and contamination, which matter in Uber's two-sided market, and how you'd read the result and decide. This round tests experimentation rigor, so name the pitfalls before you're asked.

Rider conversion dropped 3% in one city last week — how do you investigate?

Confirm the data is real first, then segment: platform, rider type, funnel step, time of day, and supply conditions (driver availability, ETAs, surge). Form a hypothesis tied to the marketplace — e.g., a supply shortage raising wait times — and test it. End with a recommendation. Uber wants product sense plus analytical rigor, not just a list of cuts.

An experiment shows a tiny but significant lift — would you ship it?

Separate statistical significance from practical significance. Acknowledge the result is real but weigh effect size against engineering cost, guardrail movements, and long-term effects. Mention novelty effects and whether the lift holds across segments. Showing judgment beyond the p-value is exactly what Uber's experimentation culture values.

What Uber looks for

Strong SQL and the ability to compute the right metrics
Solid statistics and experimentation — A/B design, significance, power
Product sense for a two-sided marketplace and its tradeoffs
Judgment that goes beyond statistical significance to business impact
Clear communication of technical results to non-technical partners
Red flag: treating a significant p-value as an automatic ship decision

FAQ

How much SQL does the Uber data scientist interview involve?

A fair amount. Expect at least one analytical SQL round with joins, window functions, and metric computation, plus SQL in the technical screen.

How heavy is the statistics and A/B testing focus?

Heavy. Experimentation is central to Uber's culture, so be ready to design A/B tests, reason about significance and power, and discuss pitfalls like network effects in a marketplace.

Is there coding beyond SQL?

Some lighter Python or basic algorithm/data-manipulation work can appear, but the emphasis is SQL, statistics, and product analytics rather than hard DSA puzzles.

Do I need marketplace knowledge?

Understanding two-sided marketplace dynamics — supply, demand, and how they interact — helps a lot in the product-metrics and experimentation rounds. Be ready to reason about both sides.

How long does the process take?

Typically a few weeks from screen to onsite to decision, varying by team and scheduling.

Walk into the Uber loop fluent in metrics and experiments

Designing an A/B test or diagnosing a metric drop out loud is harder than reading about it. Practice this exact interview — the SQL, the experimentation round, the product-sense case — with OfferLoop's realtime voice coach before the real thing.

Practice this interview out loud →

All Uber interviews →Uber Software Engineer interview →Uber Data Analyst interview →Uber Data Engineer interview →Uber Machine Learning Engineer interview →

OfferLoop is an independent interview-practice tool and is not affiliated with, endorsed by, or sponsored by Uber. All company names and trademarks are the property of their respective owners.

Interview formats vary by team, level and year, and this guide is compiled from general knowledge of publicly discussed hiring processes — treat it as preparation material, not an official description of Uber's current process.