Company interview guide

Intuit Machine Learning Interview Guide (2026)

A data-science-heavy loop built around the craft demo: model a dataset, then defend it to a panel.

Intuit's machine learning interview loop has 5 reported stages (Recruiter call and hiring manager chat, Technical screen, Craft demo: case exercise and panel presentation, Assessor interviews: ML, statistics and coding depth, Values and hiring manager interviews); typical length: 1-2 months.

Roles covered: Data Scientist and ML Engineer. The question bank holds 4 questions reported from Intuit interviews, 1 of them free to solve in your browser.

Try a free question

Interview process length
1-2 months
Stages
5
Reports linked to bank questions
4
Bank report dates
Aug 2016–Sep 2026
Questions in bank
4 (1 free)
Last reviewed
October 2026

How the Intuit interview loop works

A recruiter call, sometimes followed by a hiring manager chat, then one technical screen of about 60 minutes. The final round is a single day of four to six hours, usually virtual: the craft demo comes first, followed by assessor interviews and a values or manager conversation. Some teams send the craft case as a take-home days ahead; others give you 60-90 minutes to solve it at the start of the final day. Reports since 2024 also mention a past-project presentation in place of, or alongside, the case. Several candidates were offered a level below the one they applied for, and slow scheduling between stages is a common complaint.

The Intuit interview process, stage by stage

01Recruiter call and hiring manager chat

Short recruiter or HR call; on some teams a 30-60 min hiring manager conversation follows

A screen on background, interest in the role, level and logistics such as work authorization. Several teams add a call with the hiring manager or team lead before any technical round, which walks through your experience and occasionally includes light modelling questions.

What it tests

  • How your past work maps to the team's problems
  • Seniority and level fit
  • Logistics: location, timing, work authorization

How to prepare

  • Ask the recruiter whether the craft case is a take-home or a timed exercise on the day, and how long you get.
  • Confirm the level you are being considered for early; down-levelling after the final round is reported several times.
  • Prepare a two-minute summary of one project with clear business impact.

A recruiter or HR call opens the loop in 26 candidate reports (12 since 2024); 8 also describe an early hiring manager chat. An Intuit blog interview account also describes a recruiter screen.

02Technical screen

Usually one 60 min video round: live SQL plus Python or pandas, with basic ML or statistics questions; some teams use an online assessment or an algorithm-style coding question

The main filter before the final day. Most screens pair SQL (joins, window functions, building KPIs from sample tables) with Python or pandas data manipulation, sometimes in a notebook with a provided CSV, plus a few ML, probability or product questions. A minority of data science screens are general algorithm coding at medium difficulty, which candidates often found surprising.

What it tests

  • SQL on event or business tables: joins, windows, gaps and intervals, metric definitions
  • Cleaning and reshaping a dataset in pandas under time pressure
  • ML basics: overfitting, precision and recall, how tree ensembles work
  • Probability questions such as conditional probability or sampling from a distribution
  • Occasionally an algorithm problem, including dynamic programming

How to prepare

  • Drill SQL with window functions and interval logic until you can write it without looking anything up.
  • Practise pandas groupby, merge and string handling on a raw CSV in a notebook, not only in SQL.
  • Keep one or two medium algorithm problems warm in case your screen is coding-led.
  • Explain the metric you would build before you write the query; interviewers ask why.

Free reported questions for this stage

34 candidate reports (13 since 2024) describe it; SQL named in about 20, Python or pandas in 6, algorithm coding in 7. An Intuit blog account describes an assessor phone screen, but as a skills chat, not live coding.

03Craft demo: case exercise and panel presentation

First part of the final day.

Either a 60-90 min timed exercise on a provided dataset or a take-home case sent days ahead, then about 60 min presenting to a panel of about four with Q&A

Intuit's signature stage. You introduce yourself, share a professional achievement, then present your work on a case, typically a classification or prediction task where you define the target, engineer features, model and explain results and business impact. The panel asks hard follow-ups on target definition, data edge cases and metric choice. Take-home versions range from a few hours to a week of work; one 2026 report describes presenting a deep past ML project instead of a fresh case.

What it tests

  • Framing a vague business problem: target definition, metric choice, baseline
  • End-to-end modelling on messy tabular data within a fixed time
  • Feature engineering and handling of leakage, imbalance and edge cases
  • Telling a clear data story that leads with business impact
  • Defending decisions under questioning from a mixed panel

How to prepare

  • Rehearse a full 90-minute dataset-to-slides run: clean, baseline, one better model, three findings.
  • Lead with the business question and impact; one candidate was told their deck was too academic.
  • Prepare the self-introduction and achievement slides in advance; they are part of the same hour.
  • Test your tools and environment beforehand; several candidates lost time to platform problems.

29 of 57 reports describe it (9 since 2024); about 12 timed on the day, 14 take-home. The company's hiring page describes it, but only the on-the-day version: 90 min to solve, then 60 min with 4 people.

04Assessor interviews: ML, statistics and coding depth

Two to four one-on-one or two-on-one rounds of about 45 min after the craft demo; ML engineer loops add system design

Interviewers whose work matches the role dig into your craft case and your resume. Expect fundamentals of regression, logistic regression and tree ensembles, statistics and probability, an applied case such as how you would build a fraud model, and often a SQL or coding round. ML engineer loops reported system design and data pipeline performance discussions, and some data science candidates were tested on software engineering and complexity.

What it tests

  • Linear and logistic regression, trees, random forests and boosting, and when each fails
  • Probability and statistics puzzles, from binomial games to streaming medians
  • An applied ML case such as a fraud or risk model, including deployment and A/B testing
  • SQL or Python coding, sometimes a small NLP or text-processing task without libraries
  • For ML engineers: system design, pipeline latency and reliability

How to prepare

  • Be able to explain XGBoost and random forest versus gradient boosting from first principles in a few minutes.
  • Expect follow-ups on your craft case in every round, so know why you made each choice.
  • Review time complexity; some data science candidates were asked about algorithms on large data.
  • Tie each modelling answer to how it would be deployed, monitored and tested.

21 candidate reports describe technical one-on-ones after the craft demo (6 since 2024). The company's hiring page describes one 45 min interview with 2 assessors; reports often mention two to four rounds.

05Values and hiring manager interviews

A 30 min values interview with a cross-functional colleague plus a 30-60 min manager conversation; sometimes a stakeholder or group session

The close of the final day. A colleague from another function asks how you live Intuit's operating values, such as customer obsession and stronger together, and a people manager asks about motivation, successes and lessons. Some loops add a stakeholder partnership round; a few candidates found the manager round turned technical.

What it tests

  • Stories on influencing technical and non-technical stakeholders
  • Prioritization and handling a skeptical partner
  • Alignment with Intuit's operating values
  • Ownership, impact and what you learned from failures

How to prepare

  • Read Intuit's five operating values and map one story to each.
  • Prepare a stakeholder conflict story with a measurable outcome.
  • Do not treat the manager round as purely behavioral; be ready for technical follow-ups.

12 candidate reports describe a behavioral, values or manager round in the final (7 since 2024); the company's hiring page describes both a values interview and a manager interview.

Questions and themes reported most often

  • Model a provided dataset, then present it to the panel (29 reports, craft demo: case exercise and panel presentation)
  • Live SQL: joins, window functions and KPIs from sample tables (23 reports, technical screen)
  • Walk through a past project and defend features, metrics and edge cases (10 reports, craft demo: case exercise and panel presentation)
  • Algorithm-style coding in a data science loop (7 reports, technical screen)
  • Product or business case, often with A/B testing (7 reports, assessor interviews: ml, statistics and coding depth)
  • Tree ensembles: boosting, XGBoost, random forest versus gradient boosting (6 reports, assessor interviews: ml, statistics and coding depth)
  • Python and pandas manipulation of a raw dataset (5 reports, technical screen)
  • Linear and logistic regression: assumptions, limits, comparison with trees (4 reports, assessor interviews: ml, statistics and coding depth)

What Intuit asks about

  • Craft demo & modelling case28%
  • SQL & pandas20%
  • ML & statistics concepts20%
  • Behavioral & values12%
  • General coding12%
  • Product & business case8%

Intuit interview prep checklist

  • Built and presented a classification model on a new tabular dataset in under 90 minutes
  • Have a self-introduction and one achievement slide ready for the craft demo
  • Can write SQL with window functions and interval logic without help
  • Cleaned and summarized a messy CSV in pandas against the clock
  • Can explain boosting, random forests and regression assumptions clearly
  • Mapped a story to each of Intuit's operating values

Tips

  • Treat the craft demo as the deciding round: rehearse a timed dataset-to-presentation run at least twice.
  • Ask early whether your craft case is a take-home or a timed exercise, and plan prep time to match.
  • Lead every presentation with the business problem and impact, then the modelling.
  • Get fluent in SQL window functions and pandas; together they cover most technical screens.
  • Know classical ML deeply, especially regression, logistic regression and boosted trees, more than the newest methods.
  • Keep a few medium algorithm problems fresh; some data science screens are coding-led.
  • Prepare stories for each of Intuit's operating values, with stakeholder influence as a priority.
  • Expect slow scheduling and follow up politely; confirm your target level before the final day.

See the rest of the Intuit loop

The later stages, the questions reported for them and how to prepare are in Pro.

Unlock with Pro

Practice the Intuit loop

Work the stages in order: every reported question for this company, grouped by round, with solutions in the app.

Intuit interview FAQ

How many interview rounds does Intuit have for ML roles?

Candidates report 5 stages: Recruiter call and hiring manager chat, Technical screen, Craft demo: case exercise and panel presentation, Assessor interviews: ML, statistics and coding depth and Values and hiring manager interviews. A recruiter call, sometimes followed by a hiring manager chat, then one technical screen of about 60 minutes. The final round is a single day of four to six hours, usually virtual: the craft demo comes first, followed by assessor interviews and a values or manager conversation. Some teams send the craft case as a take-home days ahead; others give you 60-90 minutes to solve it at the start of the final day. Reports since 2024 also mention a past-project presentation in place of, or alongside, the case. Several candidates were offered a level below the one they applied for, and slow scheduling between stages is a common complaint.

How long does the Intuit ML interview process take?

Reported timelines run 1-2 months (Roughly 1-2 months).

What does Intuit ask in ML interviews?

By reported question mix: Craft demo & modelling case 28%; SQL & pandas 20%; ML & statistics concepts 20%; Behavioral & values 12%; General coding 12%; Product & business case 8%. The most repeated themes: Model a provided dataset, then present it to the panel; Live SQL: joins, window functions and KPIs from sample tables; Walk through a past project and defend features, metrics and edge cases.

How should I prepare for the Intuit ML interview?

Treat the craft demo as the deciding round: rehearse a timed dataset-to-presentation run at least twice. Ask early whether your craft case is a take-home or a timed exercise, and plan prep time to match. Lead every presentation with the business problem and impact, then the modelling.

Are there free Intuit interview questions?

Yes. LeetML has 4 questions reported from Intuit interviews, and 1 of them are free to solve in the browser; they are listed by stage above.

Related interview guides

Official sources