>
Intuit
Intuit Data Scientist Interview Guide 2026
Complete Intuit Data Scientist interview guide. Understand the assessment, the signature Craft Demonstration, ML and analytics panels, and how Intuit's AI-driven expert platform strategy shapes hiring.
5 min read
Updated Sep 2026
67+ practice questions
67+
Practice Questions5
Rounds6
Categories5 min
ReadTL;DR
Intuit's Data Scientist process stands out for its Craft Demonstration, a presentation round where you walk a panel through a past project or a prepared case, showing your end-to-end craft. The typical loop is a recruiter screen, a technical screen or online assessment covering Python, SQL, statistics, and ML, then a virtual onsite combining the craft demo with technical panels and behavioral interviews built on Intuit's values. Products like TurboTax, QuickBooks, and Credit Karma give the questions a fintech flavor, document extraction, risk models, personalization, and the AI features in Intuit Assist. The technical bar is moderate and applied, and communication carries unusual weight because of the demo format. Expect 4 to 8 weeks end to end.
4-8 weeks
67+ questions
Sample Questions
67+ in practice bank
Detect fraudulent QuickBooks payment transactions
Frame a fraud model under extreme class imbalance, discuss labels and feedback delay, cost-sensitive thresholds, and how you monitor drift as fraud adapts.
Evaluate an AI assistant answering tax questions
Design an evaluation for LLM-generated tax guidance, accuracy against ground truth, hallucination detection, escalation to human experts, and metrics that gate rollout.
SQL for subscription churn cohorts
From subscription events, compute monthly churn by acquisition cohort and plan tier, using window functions to identify at-risk cohorts after a price change.
A/B test for a new onboarding flow in TurboTax
Design the test around a strongly seasonal product, choose completion and accuracy guardrails, and discuss how tax season timing constrains experiment windows.
Explain calibration and why it matters for credit offers
Distinguish discrimination from calibration, show how a well-ranked but miscalibrated model misprices risk, and describe recalibration methods.
Tell me about a time you were transparent about a model's weakness
Integrity without compromise is a scored value. Describe surfacing a limitation that cost you short-term, and how honest framing changed the decision.
About the Interview Process
Intuit runs a structured, values-forward process. The assessment and technical screen filter fundamentals, then the onsite is anchored by the Craft Demonstration, unique among big tech peers, where a panel of scientists and managers evaluates how you frame problems, make decisions, and communicate. Technical panels dig into ML and statistics behind your demo and beyond, and behavioral rounds map explicitly to Intuit's published values.
Recruiter Screen
Background and role overview. Intuit has DS roles across TurboTax, QuickBooks, Credit Karma, Mailchimp, and platform AI teams, clarify which product area and whether the role tilts analytics or ML.
Technical Screen / Assessment
Online assessment or live screen with Python data manipulation, SQL, probability and statistics, and applied ML questions. Difficulty is moderate, breadth and correctness matter most.
Craft Demonstration
You present a past project or prepared case to a panel, then defend methodology, alternatives, and impact under questioning. Storytelling quality, technical depth, and honest handling of limitations are all scored.
Panel: ML & Analytics
Applied ML and statistics beyond your demo, modeling cases from Intuit's domain such as fraud detection or document field extraction, plus evaluation, deployment, and monitoring discussion.
Panel: Behavioral & Values
Structured questions against Intuit's values with managers and cross-functional partners. Expect customer-obsession stories, integrity moments, and collaboration under disagreement.
Timeline
4 to 8 weeks. The craft demo adds scheduling coordination, panels need multiple attendees.
Tips
Structure your demo deck for interruption, panels ask questions throughout, so make each slide self-contained.
Quantify impact in customer and business terms, refunds processed faster, chargebacks reduced, hours saved.
When you do not know an answer in panels, reason aloud from fundamentals, the culture rewards visible thinking.
Ask your recruiter exactly what demo format the team expects, past project versus provided case varies by group.
The Craft Demonstration
The craft demo is Intuit's distinctive filter and most candidates underprepare for it. The panel is evaluating craft, how you chose the problem framing, what alternatives you considered, how you validated, and whether you represent uncertainty honestly. A polished narrative about a modest project beats a scattered tour of an impressive one.
Build the presentation as a decision story, the business question, the constraint that made it hard, two or three approaches you weighed, evidence for your choice, and measured impact with a limitation you would fix next. Rehearse to fit half your slot, live questioning fills the rest.
What the technical rounds test
Intuit's domain is money movement, taxes, accounting, and credit, so applied questions concentrate on risk scoring, anomaly and fraud detection, document understanding, and personalization. Calibration, cost-sensitive evaluation, and explainability come up because model errors carry direct financial consequences for customers.
With Intuit Assist embedding GenAI across products, many teams now ask about LLM-era topics, evaluating generated answers, guarding against hallucinated financial guidance, and combining structured models with LLM interfaces. You do not need deep LLM research experience, but a considered view on evaluating AI assistants in high-stakes domains is increasingly expected.
Leveling & Compensation
| Level | Title | YoE | Total Comp (USD/yr) |
|---|---|---|---|
DS1 | Data Scientist 1 | 0-2 yrs | $130k - $195k |
DS2 | Data Scientist 2 | 2-5 yrs | $160k - $250k |
Sr DS | Senior Data Scientist | 5-9 yrs | $190k - $300k |
Staff DS | Staff Data Scientist | 8+ yrs | $230k - $370k |
Data Scientist 1
Delivers analyses and model iterations with mentorship. Solid fundamentals and growing product context.
Data Scientist 2
Owns models or measurement for a product area. Ships through experiments, partners with PM and engineering, and presents to leadership.
Senior Data Scientist
Technical lead for significant modeling initiatives. Sets methodology, mentors, and drives multi-quarter roadmaps.
Staff Data Scientist
Shapes DS direction across teams, owns architecture of the modeling stack for a domain, and influences org-level strategy.
How to Stand Out
Behavioral Focus Areas
Customer obsession: Intuit's design-for-delight culture expects customer empathy from scientists
Integrity without compromise: honesty about model limitations, essential in financial products
Courage: proposing bold ideas and defending sound methodology under challenge
We care and give back: collaboration and mentorship stories resonate
Ownership: driving projects from ambiguity to shipped impact
1.
Treat the Craft Demonstration as the round to win. Build a clear narrative, problem, approach, validation, impact, and rehearse it aloud with a strict time limit.
2.
Choose a demo project with real trade-offs you can defend, panels probe alternatives you rejected and why.
3.
Refresh applied ML breadth, tree ensembles, regularization, evaluation under imbalance, and calibration, common in risk and personalization work.
4.
Practice explaining a model's errors, financial products care about failure modes more than average accuracy.
5.
Review basic experimentation, Intuit ships through A/B tests and asks standard design questions.
6.
Connect your motivation to Intuit's customers, small businesses and taxpayers, values questions are asked earnestly.
Recommended Resources
FAQ
What exactly is the Craft Demonstration?
A panel presentation, usually 45 to 60 minutes with questions, where you walk through a project demonstrating your end-to-end craft. Some teams ask for a past project, others provide a case in advance. It is the highest-weighted round on most Intuit DS loops, prepare it like a conference talk.
Can I use a work project under NDA for the demo?
Yes, with sensible abstraction, rename entities, index the numbers, and keep the methodology story intact. Panels care about your decisions and reasoning, not proprietary specifics, and they respect discretion.
How ML-heavy is the role versus analytics?
Intuit has both flavors. Product DS roles lean analytics and experimentation, AI and risk teams lean ML engineering-adjacent. The job description keywords, and a direct question to the recruiter, will tell you which panels to expect.
How does Credit Karma fit in?
Credit Karma operates semi-independently with its own DS org and a similar but separately run process, heavier on credit modeling and personalization. If your role sits there, expect more risk-model depth and the same craft-demo spirit.
Comments
