Uber is a multi-national ride-hailing company with massive operations in over 785 cities worldwide. Their services range from ride-hailing and food delivery to logistics and micro-mobility. Uber’s aim is to:

Bring the future closer to its customers with self-driving technology and urban air transport, helping people order food quickly and affordably, removing barriers to healthcare, creating new freight-booking solutions, and helping companies provide a seamless employee travel experience.

To achieve this feat, Uber has slowly been incorporating data science and analytics in almost every department and services — such as risk management, marketing, and policy implementation.

The Data Science Role at Uber

The role of a data scientist at Uber varies across specific teams. Your role as a data scientist will be heavily determined by the team you are applying for. The data science role generally covers basic business analytics, modeling, machine learning, and deep learning implementation. Uber is a large company that has data science teams working in safety and insurance, rides, risk, platform, marketing science, policy, and Uber Eats.

Required Skills

The requirements for each job depends on the department. However, Uber generally prefers to hire qualified candidates with at least three years of experience unless for an associate position.

The basic requirements for hiring are:

  • Ph.D., M.Sc., B.Sc. or B.A. degree in Statistics, Mathematics, Economics, Operations Research, Computer Science, Physics or any other related quantitative field. (Advanced degrees are an advantage).
  • 3 years’ experience or more (at least 10 years for senior data scientist) in experimental design (A/B testing preferable), exploratory data analysis, statistical analysis, and machine learning model development.
  • Proficiency in at least one programming languages such as Python, Java, R, or SQL.
  • Experience building pipelines and ETLs that transform structured and unstructured data sets in large-scale, complex datasets (Hadoop, Hive, Vertica, Presto) into actionable data models.

What kind of data science role?

There are a large variety of data science teams at Uber that work spread out across different divisions of the company. The title “data scientist” at Uber falls under these two teams — Data Science and Data Science & Analytics.

Depending on the teams, their functions may include:

  • Safety and Insurance: Implementing machine learning algorithms, and optimizing safety policies to help reduce safety related incidents for customers.
  • Core Client: Forecasting and automating every aspect of Uber’s core ride-sharing products.
  • Risk: Developing machine learning models and strategies to curb and manage market place abuse and payment fraud.
  • Research: Conducting in-house research to better understand the market environment and improve product.
  • Marketing Science: Applying statistical modeling, machine learning, or data mining techniques in providing valuable insights across Uber’s global marketing efforts.

The Interview Process

The interview process starts with an initial screen with a recruiter or hiring manager that lasts for 15 and 30 minutes respectively. This is followed by an Uber take-home challenge. The take-home assignment covers SQL questions, experimental/business intelligence questions, and data analysis questions. Then a 45-minute technical phone screening follows the take-home challenge. After the technical screen is the on-site interview panel of five different interviewers.

Looking for more help with SQL interview questions? Take a look at our ultimate guide here.

Initial Screen

This is a phone call interview with a hiring manager or recruiter after your application submission. This interview is about assessing the job role, the team, and your general background with potential light technical interview. Expect questions about your past experience and how it could apply to Uber. The hiring manager may also ask more high-level technical questions such as:

  • What is marketing attribution?
  • What metrics would you use to measure a model’s effectiveness?
  • How would you explain a p-value to a non-technical person?

The hiring manager is generally looking for red flags. Make sure to review general modeling and analytics concepts and practice communicating technical concepts and projects.

The Take-Home Challenge

source: Wired

After completing the initial phone screening, you will receive a take-home challenge that you will have one week to complete. The take-home assignment comprises of three sections:

  • Sql and Analytics: You’ll be given an example Uber problem with a schema. The question will ask to write SQL to solve various analytics problems.
  • Qualitative section: General questions on metric evaluation and experimental design.
  • Modeling: An applied predictive modeling exercise.

Note that this take-home challenge has been generally standardized by Uber. Depending on the team however, they may add in changes to the original take-home challenge specific to the team.

Practice a SQL interview question:

Try our built-in SQL editor!
Given the tables above, select the top 3 departments by the highest percentage of employees making over 100K in salary and have at least 10 employees.

Technical Screen

The next step in the process is the technical interview phone interview with a data scientist. Most of the time, the questions asked in this interview are Uber-related case studies looking for an open-ended response. The goal here is to test your critical thinking and problem-solving ability. Expect to receive machine learning problems like feature selection and model building, with a focus on real-life Uber problems. If the role is more analytics focused you can expect a product based question as well.

Example Questions:

  • What are performance metrics for evaluating various Uber services?
  • How do you investigate that a certain trend in the distribution is due to anomaly?
  • What problems have you faced with supervised machine learning and how do you overcome them?
  • How would you predict ride requests? How would you evaluate the estimated time to arrival algorithm in Uber?

Onsite Interview

The next step after passing the technical screen is the on-site interview. The on-site interview consists of 5 or 6 rounds of 45-minutes each. This is a full-day interview involving whiteboard coding, project discussion with team managers and data scientists, business case studies, and statistical concept discussions.

The panel generally looks like:

  • A one-on-one interview with a data scientist. You will be given some open-ended business intelligence and analytics problem with a statistics and probability question as well.
  • A behavioral interview with a product manager.
  • A hiring manager interview going over a deep-dive into Uber and a discussion about the team. Make sure to ask thoughtful questions here.
  • A technical machine learning interview with a data scientist. This interview goes over modeling concepts and machine learning design questions.
  • A 45-minute long interview with a data scientist that involves coding in SQL or algorithms. SQL if the role is in the analytics division and algorithms if it’s in the machine learning division.

Remember that ultimate goal is to assess how you can apply data science concepts to Uber-related specific business problems. Brush up on knowledge of statistics and probability, A/B testing and experimental design, and modeling concepts.

In terms of technical knowledge, remember to practice coding and SQL exercises and problems that can be found on Interview Query. Practicing more of these questions can help with getting passed the baseline level in technical skill.

Sample Uber Data Science Interview Questions

  • Describe linear regression to a child, to a first-year college student, and to a seasoned mathematician.
  • Let’s say we launch a new Uber Eats feature. What would you choose as the key metric?
  • How would you design an incentive scheme for drivers such that they would more likely go into city areas where demand is high?
  • Given a random Bernoulli trial generator, write a function to return a value sampled from a normal distribution.
  • What metrics would you use to track Uber’s strategy of using paid advertising to acquire customers’ works? How would you figure out an acceptable cost of customer acquisition?
  • What are the costs of having a fleet of vehicles take Google street view photos of every major city in the US every day?
  • Build a text wrapper. For example, split a long sentence by some character limit only at the spaces.
  • Write a production code to find all combinations of numbers in a list that sum up to 8.
  • What is the difference between MLE and MAP?
  • What are the assumptions of linear regression?
  • What do nested SELECT and WITH do in SQL?
  • What algorithm would you use to predict if a driver will accept a ride request or not? What features would you use?

Want more interview questions with solutions from Uber? Find more on Interview Query.