Intuit Data Scientist interview typically runs 4 rounds: technical screens, take-home assignment, onsite loop, offer. The process takes about 1-2 weeks and is notably intense, with a heavy time commitment.
$165K
Avg. Base Comp
$280K
Avg. Total Comp
3
Typical Rounds
3-5 weeks
Process Length
We’ve seen Intuit evaluate for more than polished answers; they want candidates who can hold a product and an experiment in their head at the same time. A recurring theme in candidate reports is the intensity of the probing once you get past the screens: interviewers quickly moved from a take-home into old projects, recommendation systems, ML choices like XGBoost, and then a full QuickBooks A/B test scenario. The signal is clear: they care about how you reason through tradeoffs, metrics, and failure modes, not just whether you can name the right method.
Another pattern we’ve noticed is how much weight they place on specificity. One candidate said the team expected a week-long take-home and still pressed on the depth of the analysis, which included prediction, experiment design, EDA on a 250k-row dataset, recommendations, and a slide deck. That tells us Intuit is looking for people who can turn analysis into a decision-ready narrative. When candidates got vague, the follow-up questions got sharper, especially around primary vs. guardrail metrics, hypothesis framing, and what could go wrong.
The non-obvious part here is that the bar seems tied to calibration as much as capability. Multiple candidates reported that interview performance influenced compensation placement, and that the questioning was intentionally exhaustive. In practice, that means the strongest candidates are the ones who can stay crisp under pressure and defend every assumption without drifting into generic explanations. At Intuit, the difference between “good” and “hire” often comes down to whether your thinking feels operational enough to ship inside a product team.
Synthetized from 1 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 Intuit process.
I got through all the technical screens and made it to the onsite loop, but ultimately decided not to move forward because of layoff and restructuring rumors and because the role expectations felt really vague, with performance reviews starting at three months.
The technical screens used CoderPad and also Glider, which is an AI proctoring tool that tracks eye movement and body language on top of screen sharing. That was more intense than anything else I encountered across my interviews.
For the loop, Intuit gave a take-home assignment that they expected candidates to spend at least a week on. I took two days off and worked on it, and they said that wasn't sufficient. The assignment had five different analyses:
After the take-home, the loop rounds were called "ask anything" rounds. One hour was focused on the assignment, and the other three hours were two data scientists (one staff, one senior manager) jumping from topic to topic asking whatever they wanted. They went into my old projects, recommendation systems, ML algorithms like XGBoost, and then ran through a full A/B test scenario on a QuickBooks product. For that scenario, they expected you to cover initial considerations, primary metrics, secondary metrics, drivers, guardrail metrics, hypothesis, what could go wrong, and scaling of A/B testing (sequential vs. parallel) all in about 20 minutes. If you were vague on anything, they'd dig deeper right there.
I heard from other people that the drilling is partly by design to calibrate the offer, since Intuit gives a wide salary range and uses interview performance to decide where in that range to land you.
Intuit requires a serious time commitment before you even get an offer. If you're going to do the take-home, plan for a full week, not a couple of days. And for the loop, be ready to go deep on A/B testing end to end, including sequential vs. parallel experiment scaling, because they will find the gaps and push on them.
Prep tip from this candidate
Intuit's 'ask anything' loop rounds cover A/B testing in serious depth, including sequential vs. parallel scaling, guardrail metrics, and QuickBooks-specific product scenarios, all at speed. Budget a full week for the take-home assignment, which includes building a prediction model, designing an experiment, doing EDA on a 250k-row dataset, and producing a 15-20 slide deck.
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.
| Question | |
|---|---|
| Digitizing Student Test Scores | |
| Running Dog | |
| Overfit Avoidance | |
| Approval Drop | |
| Above Average Product Prices | |
| Client Solution Pushback | |
| Why Do You Want to Work With Us | |
| Your Strengths and Weaknesses | |
| Empty Neighborhoods | |
| 2nd Highest Salary | |
| Rolling Bank Transactions | |
| Top Three Salaries | |
| Customer Orders | |
| Comments Histogram | |
| Closest SAT Scores | |
| Subscription Overlap | |
| Merge Sorted Lists | |
| Upsell Transactions | |
| Monthly Customer Report | |
| First Touch Attribution | |
| Experiment Validity | |
| First to Six | |
| Last Transaction | |
| Compute Deviation | |
| Download Facts | |
| Button AB Test | |
| Bank Fraud Model | |
| Top 5 Turnover Risk | |
| Top 3 Users |
Synthesized from candidate reports. Individual experiences may vary.
Candidates go through technical screening rounds using CoderPad and Glider. The screens are intensive and include live coding plus proctored assessment conditions, with Glider monitoring screen sharing and even eye movement/body language.
Intuit gives a substantial take-home project that candidates are expected to spend at least a week on. The assignment includes building a prediction algorithm, designing an A/B test, exploratory data analysis on a ~250k-row dummy dataset, proposing solutions from experiment output, and preparing a 15-20 slide presentation of findings.
The onsite loop is made up of several 'ask anything' rounds. One hour is focused on the take-home, and the remaining three hours involve two data scientists, including a staff-level and a senior manager, asking broad technical questions across prior projects, recommendation systems, ML algorithms like XGBoost, and an end-to-end A/B test scenario on a QuickBooks product.