Google Data Scientist interviews typically run 4–5 rounds: recruiter screen, two screening rounds, two technical rounds, and a Googliness behavioral round. The process spans roughly 4–8 weeks and is distinguished by its balanced mix of SQL, product sense, and culture-fit evaluation.
$123K
Avg. Base Comp
$330K
Avg. Total Comp
4-5
Typical Rounds
4-6 weeks
Process Length
What stands out most across Google's data scientist interviews is how consistently the process blends technical execution with product reasoning — and how candidates who treat these as separate preparation tracks tend to struggle. We've seen candidates clear the SQL and Python rounds comfortably, then get tripped up when asked to frame a modeling problem for YouTube Shorts or articulate how they'd improve Google Maps. The expectation isn't just that you can write a window function — it's that you can explain why a metric matters and what success looks like for a real product.
A recurring theme in candidate reports is the weight placed on experimentation fluency. It's not enough to know what an A/B test is. Multiple candidates encountered questions about sanity checks before launch — sample ratio mismatch, randomization unit validity, pre-experiment balance across dimensions like geography and device type. This is the kind of question that separates candidates with real experimentation experience from those who've only read about it. The probability and statistics questions we see in the question bank — coin bias simulations, Bayesian reasoning, non-normal A/B testing — reinforce that Google wants you to think probabilistically, not just procedurally.
The 'Googliness' round is easy to underestimate. Candidates who received offers consistently noted that behavioral stories needed to be specific and tied to collaboration, difficult decisions, and leadership moments — not vague summaries of past roles. The technical bar here is real but manageable; what separates offers from near-misses is often how clearly a candidate communicates their reasoning, both in code and in conversation.
Synthetized from 3 candidates reports by our editorial team.
Had an interview recently?
Share your experience. Unlock the full guide.
Real interview reports from people who went through the Google process.
I got into Google's process without an HR screening call at all. They just emailed me directly saying I'd been selected for a technical screen. That was a nice surprise. The process ended up being two technical rounds, then an on-site that mirrored those same two rounds plus an HR round.
The first technical screen covered AB testing and SQL — it was the heavier round in terms of breadth. The second was more of an ML case study where they gave me a scenario and asked how I'd approach it, including what models I'd use. It was high-level, not a coding round. More like, here's a situation, walk us through your thinking.
The on-site was basically the same two formats I'd already done in the technical screen, so having gone through those earlier I kind of knew what to expect. They added an HR round on top. The recruiters did a really great job preparing me — whatever they shared with me matched pretty closely to what actually happened, so nothing really threw me off.
For the AB testing questions, they really wanted to see that you understood the full lifecycle — not just setting up the test but interpreting results, handling edge cases like novelty effects, and knowing when a test result might be misleading. They asked things like how you'd handle a situation where your metric moved but you weren't sure if it was a real effect. For SQL, it wasn't just syntax — they wanted to see how you'd structure a query to answer a real product question, so thinking out loud about your logic mattered as much as getting the right answer.
The ML case study round felt more like a product sense conversation with a modeling layer on top. They gave me a scenario tied to a Google product and asked how I'd frame the problem, what data I'd want, and what model I'd reach for. They pushed back on my answers to see how I'd defend my choices, so being able to explain your reasoning clearly is really important there.
I didn't get an offer. The feedback I got was that it came down to me and another candidate, and the other person just gave stronger signals for what the specific team was looking for. It was a close call, which honestly made it harder to hear, but the process felt fair and the interviewers were engaged throughout.
One thing I'd pass along: treat the technical screen as practice for the on-site, not just a gate. The formats are the same, so if you make it through the screen, you'll have a real preview of what's coming. Also, don't underestimate the HR round — it's not just a formality. They're assessing whether you can communicate your work clearly to non-technical stakeholders, so practice telling the story of your past projects in plain language.
Prep tip from this candidate
The Google DS process here had two core round types repeated across both the technical screen and the on-site: an Applied Analysis round (AB testing plus SQL) and an ML case study round (high-level scenario, model selection). Nail both formats early because the on-site is essentially a repeat of the screen.
Share your own interview experience to unlock all reports, or subscribe for full access.
Sourced from candidate reports and verified by our team.
Topics based on recent interview experiences.
Featured question at Google
Write a query that returns all neighborhoods that have 0 users.
| Question | |
|---|---|
| 2nd Highest Salary | |
| Top Three Salaries | |
| First Touch Attribution | |
| Merge Sorted Lists | |
| First to Six | |
| Last Transaction | |
| Experiment Validity | |
| 500 Cards | |
| String Shift | |
| Minimum Change | |
| Lazy Raters | |
| Jars and Coins | |
| Bucket Test Scores | |
| Complete Addresses | |
| Find Bigrams | |
| Network Experiment Design | |
| Impression Reach | |
| The Brackets Problem | |
| Friendship Timeline | |
| Lifetime Plays | |
| Detecting ECG Tachycardia Runs | |
| Reducing Error Margin | |
| Good Grades and Favorite Colors | |
| Button AB Test | |
| Amateur Performance | |
| Estimated Rounds | |
| Distribution of 2X - Y | |
| Fair Coin | |
| Nearest Common Ancestor |
Synthesized from candidate reports. Individual experiences may vary.
Initial contact typically happens about two weeks after applying online. The recruiter covers your past experience, role expectations, and logistics like process timeline and compensation. This round may also include a light technical or Bayesian statistics question, making it slightly more rigorous than a standard recruiter call.
Focuses on SQL and Python fundamentals. SQL questions typically involve window functions such as calculating a 7-day rolling average of daily active users. Python questions are practical rather than algorithmic, such as filtering and transforming lists of dictionaries. An experimentation question is also common, asking you to walk through sanity checks before launching an A/B test.
Centers on product sense and ML fundamentals. You may be asked to improve a Google product like Maps by identifying user segments, prioritizing pain points, and proposing features with success metrics. A high-level overview of the ML model lifecycle — from data collection through deployment — is also covered, typically in the last 10-15 minutes.
Consists of additional technical rounds covering SQL, Python, statistical simulation, and applied ML problem framing. Coding is done in Google's own document environment rather than a third-party platform. Questions span topics like simulating biased coin tosses, string manipulation, and building end-to-end predictive models for products like YouTube Shorts.
Explicitly evaluates culture fit and work style aligned with Google's values. Interviewers look for strong examples of leadership, handling difficult decisions, and cross-functional collaboration. Having well-prepared behavioral stories is essential, as this round is a distinct and weighted part of the loop.