Google Data Scientist Interview Guide

Google's data scientist interview has fewer rounds than most big tech data science loops, so each technical round covers fewer topics in more detail. If you mention a t-test, expect to compute its test statistic by hand. A hiring committee decides your offer and level from your interviewers' written feedback, which makes the structure of your answer as important as whether it's right.

This guide breaks down each stage of the Google data scientist interview process, what interviewers look for, and how to prepare with real example questions, actionable tips, and resources.

Google data scientist interview process

Google's data scientist interview process is team-independent until team matching, though the rounds you get and the depth of statistics questions can vary by track and organization. Interviewers draw questions from banks that Google refreshes regularly.

Google limits most loops to about 4 interviews, a cap it set after finding that more interviews rarely changed the hiring decision.

Here's what the interview process can look like:

• Recruiter screen: Remote call covering your background, motivation, and fit for the target team
• Technical screen: Remote SQL, data manipulation, and domain questions in a shared Google Doc
• Onsite loop (3-4 rounds): Measurement and modeling concepts, experimentation and applied analysis, Googleyness and leadership, and a SQL and machine learning round at some teams
• Hiring committee review: Googlers outside your loop review interviewer feedback and decide on the offer and level
• Team matching: Team matching can happen before or after hiring committee review, so ask your recruiter which order your process follows. When you match first, the hiring manager's support goes into the packet the committee reviews.

Google Business vs. Product vs. Research data scientist interviews

Google's Business, Product, and Research data scientist interviews follow the same core structure, and each track shifts the focus of the case and statistics rounds. Here's how the tracks compare:

Recruiter screen

The Google data scientist recruiter screen covers your background, experience, and interest in the specific team you're applying to. Hiring managers can give recruiters role-specific questions, so expect domain-relevant prompts at this stage.

Use the call to ask which rounds your loop includes and what each one covers. Recruiters may also share round-by-round feedback later in the process.

Google may also send you the Google Hiring Assessment, an online workstyle questionnaire, before or alongside the technical screen.

Interviewers look for:

• Relevant experience: Whether your background matches the team's domain and technical needs
• Domain awareness: Your familiarity with the product area, the team's priorities, and its data challenges
• Motivation: A specific reason you want to work on this team at Google
• Communication: Your ability to explain your experience clearly and concisely

Sample questions

Here are some real interview questions reported by candidates:

• Why do you want to work on this team?
• What are the biggest challenges when working with [domain] data?
• Talk about your experience working cross-functionally with engineers.
• What interests you about this product area, and what would you focus on first?

Technical screen

The Google data scientist technical screen is a remote interview with a data scientist or hiring manager that determines whether you advance to the onsite. Expect a mix of SQL, data manipulation, and schema design questions, plus one question tied to the team's domain.

Google conducts the screen in a shared Google Doc and lets you choose between Python and SQL. The SQL at this stage typically covers joins and aggregations, and window functions are more likely in the onsite.

Interviewers look for:

• SQL fluency: Your ability to write clean, correct queries quickly, including the edge cases interview questions commonly target
• Data manipulation: How you transform and aggregate raw data into the output a question asks for
• Schema design: How you translate a loosely defined product into tables and relationships
• Clarifying questions: Whether you scope the question before you start building
• Pattern recognition: Your ability to identify trends such as seasonality and explain your approach

Recently asked questions

Here are real, recent interview questions reported by candidates:

• Given a quarterly data set, find and explain the seasonality patterns.
• A company has a video-sharing app similar to YouTube. Design a schema for it, including the tables you need and how they relate, and ask clarifying questions before you build.

Measurement and modeling concepts round

The measurement and modeling concepts round in the Google data scientist onsite tests how well you understand the statistics behind the methods you use. Expect open-ended questions on distributions, sampling, and test statistics, with follow-ups on each concept you raise.

A Google L5 Product data scientist summarized the expectation for this round: "I had to know the math cold here, not just wave at the idea."

Interviewers look for:

• Statistical foundations: Your ability to explain distributions, hypothesis testing, and estimation
• Computational depth: Whether you can compute a test statistic by hand and explain each step
• Method range: Your comfort with t-tests, z-tests, maximum likelihood estimation (MLE), and sampling techniques
• Follow-up clarity: How clearly you explain a concept when interviewers question it further

Recently asked questions

Here are real, recent interview questions reported by candidates:

• You've built an arbitrary distribution. How do you sample from it?
• Walk through how you calculate a t-statistic and a z-statistic.

Experimentation and applied analysis round

Google's experimentation and applied analysis round gives data scientist candidates a metric change or open-ended product question to talk through before the interviewer shares a sample data set.

A/B testing mechanics remain core to the round, including test setup, the relevant statistics, and reliability checks. Interviewers may extend the case into causal inference when user-level randomization isn't possible.

Interviewers look for:

• Structured diagnosis: How you rule out seasonality, product launches, and logging changes before you attribute a metric movement to a cause
• Experiment design: Your ability to set up an A/B test end to end, including metric selection, randomization, sample sizing, and validity checks
• Causal inference range: Your judgment on when methods beyond A/B testing fit a scenario and why
• Statistical rigor: Whether you know the math behind each method, such as computing a two-sample test statistic for absolute vs. ratio metrics
• Business framing: How you connect the analysis to the product question and explain what the results mean for a decision

Recently asked questions

Here are real interview questions shared by Google data scientist candidates and interviewers:

• Google Meet used to be available only to G Suite users and is now widely available. How do you define success, and which metrics do you use?
• Define and walk through the full setup of an A/B test for a specific product scenario.
• You need to evaluate a change that was rolled out city by city and can't be randomized at the user level. How do you measure the impact?
• You have observational data and want to estimate a treatment effect. How do you approach it?
• Reviews on Google Maps suddenly spiked. How do you investigate what's causing the increase?
• How would you address abusive comments on YouTube?

SQL and machine learning round

Some Google data scientist onsite loops add a SQL and machine learning round that tests harder SQL than the technical screen, including subqueries, window functions, and joins across multiple tables.

Machine learning questions in this round cover fundamentals such as handling imbalanced data and choosing model evaluation metrics.

Business DS versions of this round may include a Python coding question built on a challenge the team has worked on. Other loops fold coding into the experimentation and applied analysis round.

Interviewers look for:

• Advanced SQL: Your ability to write correct queries using subqueries, window functions, and multi-table joins
• Query communication: How clearly you explain your logic as you write each query
• Imbalanced data handling: Whether you know how to train and evaluate models when one class dominates the data
• Evaluation metrics: Your judgment in choosing a metric that fits the model and the business goal

Recently asked questions

Here are real, recent interview questions reported by candidates:

• Is KNN supervised or unsupervised?
• What metrics would you use to evaluate a model?

Googleyness and leadership interview

The Google data scientist Googleyness and leadership interview is a behavioral round focused on collaboration, initiative, and adaptability. Its questions are more straightforward than the behavioral questions at Amazon or Meta, with fewer high-pressure conflict scenarios.

Interviewers look for:

• Collaboration: Whether you work effectively with engineers, product managers, and other partners outside your own function
• Initiative: Evidence that you contribute beyond your defined role without being asked
• Structured storytelling: Your ability to deliver a complete answer covering the situation, your actions, the stakes, and the results
• Adaptability: How you've handled shifting priorities or ambiguity in past roles
• Googleyness: Your comfort with autonomy and user-focused decision-making

Recently asked questions

Here are real, recent interview questions reported by candidates:

• How have you made your workplace a better place?
• Tell me about a time you adapted to changing priorities.
• Tell me about a time you contributed outside of your typical work responsibilities.

How to prepare for the Google data scientist interview

1. Practice interview-style SQL, even if you write SQL daily: SQL questions test a recurring set of edge cases and patterns that rarely appear in day-to-day analysis. Google conducts technical questions in a shared Google Doc, so practice timed SQL interview questions in a plain document without autocomplete or query execution.
2. Review test statistic math before each round: Work through t-statistic, z-statistic, and two-sample test calculations for absolute and ratio metrics on paper in the days before every round, using the statistics and experimentation lessons as a reference.
3. Present a step-by-step plan at senior levels: Interviewers expect senior candidates to outline their full approach before they analyze anything, and a clearly ordered plan supports an offer at your target level. Open each case by stating your plan out loud.
4. Explain your reasoning as you work: Practice narrating your analysis out loud with the data communication lessons.
5. Research your target team's domain: Read the team's recent blog posts, learn how the product makes money, and think through the data challenges specific to that area. For Trust & Safety roles, review the relevant product's content policies before your case rounds.
6. Generate practice cases with AI tools: Give an AI tool the job description or product area and ask for a timed, domain-specific case prompt.
7. Simulate follow-up questions with mock interviews: Practice aloud through peer and AI mock interviews to build comfort with the timed follow-up questions Google interviewers ask. For direct feedback on your case structure, work with an expert coach.

About the Google data scientist role

Google data scientists work embedded in product and engineering teams, analyzing data to identify issues and propose solutions. The role requires a working knowledge of each project's technical structure, though data scientists don't write production code or build features.

Google data scientists typically work on:

• Designing and analyzing experiments that measure product changes and user behavior
• Building statistical models to inform product and business decisions
• Defining and tracking metrics for product health, growth, and risk
• Partnering with engineers, product managers, and other stakeholders
• Segmenting and analyzing large data sets to identify trends, anomalies, and opportunities

Google data scientist experience and education requirements

Google's data scientist postings typically require a bachelor's degree in statistics, mathematics, data science, engineering, physics, economics, or another quantitative field. Required experience varies by level and posting, and most roles ask for work experience with analysis applications and coding in Python, R, or SQL.

Data Scientist III postings range from a bachelor's degree with 5 years of experience to a master's degree with 3-5 years or a PhD, depending on the team. Research data scientist roles require a master's degree or PhD, and some Ads research roles require a PhD. Business data scientist postings may also call for experience with controlled experiment design and causal inference methods.

Additional resources

• Data Science Interview course
• SQL Interview course
• Product Analytics Interview course
• Google data scientist interview questions
• Google behavioral interview questions
• Google data scientist interview experiences

FAQs about the Google data scientist interview

How long does the Google data scientist interview process take?

The Google data scientist interview process typically takes 1 to 2 months from recruiter screen to offer. The process includes a recruiter screen, a technical screen, an onsite loop, team matching, and hiring committee review, and the last 2 steps can extend the timeline after your final onsite round.

Does Google conduct data scientist interviews in person?

Google conducts most data scientist interviews virtually but is adding at least one in-person round for some roles. Confirm the format with your recruiter before your onsite.

Can you use AI tools in a Google data scientist interview?

Google doesn't allow AI tools during data scientist interviews. Google is piloting a Gemini-assisted coding round for some software engineering roles, so confirm with your recruiter before you assume any round permits AI.

Does Google offer data science internships?

Google offers internships to undergraduate and graduate students, including Business data scientist internships for PhD students. The Business data scientist PhD internship lasts 12 weeks in the US, and applications for summer internships close in the fall.

How much does a Google data scientist make?

Here are the reported compensation ranges by level for Google data scientists, according to Levels.fyi:

• L3 (Data Scientist II): ~$180K
• L4 (Data Scientist III): ~$269K
• L5 (Senior Data Scientist): ~$371K
• L6 (Staff Data Scientist): ~$478K
• L7 (Senior Staff Data Scientist): ~$702K
• L8 (Principal Data Scientist): ~$765K

Compensation combines base salary, stock, and an annual bonus, with stock making up roughly 30% of the total at L5. Google's US job postings list base salary only, from $122K at the Data Scientist III level to $278K for staff-level roles.