Intuit Interview Questions
& Process for Software Engineers

The panel in Intuit's Craft Demo withholds part of the requirements on purpose, and the first thing they write down is whether you noticed. You get a real repository, usually 24 to 48 hours ahead, so you can set it up on your own machine, and then you extend it live while two to four engineers watch and interrupt. Candidates who open the file and start implementing lose to the ones who spent five minutes asking what a guest user is allowed to see. Nobody counts how many user stories you closed. What gets written up is your layering, your tests, and the trade-offs you defended out loud. Ahead of that round sit a recruiter screen, a Glider online assessment, and a technical phone screen the process itself nicknames the Eliminator. Two to four weeks, application to offer.

The onsite has a name. Intuit calls it A4A, short for Assessing for Awesome, and it runs four to six rounds in a day. The Craft Demo anchors it, and around it sit an Assessor round, a round with someone who would be your teammate, and the hiring manager. Since 2026 an AI round shows up in most loops, less interested in whether you use a model than in how you check what it hands back. No committee sits above any of this: the same people who watched you work file their feedback into a debrief that settles the hire and the level together, so a round you treated as a warm-up costs you a vote you needed.

Two of the rounds after the Craft Demo go back to the same repository. The Assessor round opens with a data-structure problem, often under the constraint that the Java has to compile and run, moves into Spring internals and object-oriented theory, then returns to your project with a schema change or an API version to justify. The team-member round asks for another feature on top of it. Your morning's work keeps getting handed to a fresh pair of eyes, which is a decent argument for writing the tests you were tempted to skip.

Design rarely gets a box of its own. It surfaces inside that team-member round as an open discussion, or as a follow-up bolted onto a coding task, and a standalone whiteboard hour shows up mainly at senior and above. Values are read the same way, in every room rather than saved for the hiring manager, whose round is the only one that leaves your code alone.

CodingCore

Practical data-structure and algorithm problems, graded on clean, working code over trick-spotting

The Craft DemoCraft

Build a small working feature end to end, with clean, idiomatic code, sound object-oriented design, and correct edge cases

System designScale

Design a service end to end with a generic HLD playbook, reasoning about access patterns, data flow, and failure

Behavioral & valuesValues

Ownership, collaboration, and customer focus, anchored in how the team builds and operates

The algorithm bar sits at medium and stays there. Sorted by difficulty, the problems run roughly 16% easy, 64% medium, and 19% hard, and that hard tail is mostly two-dimensional and matrix DP rather than anything exotic. The surprises here are procedural rather than algorithmic. Java is the default and sometimes the requirement, the Assessor may want code that compiles and runs rather than a whiteboard sketch, and you are told to switch off Copilot and every other AI extension before you start.

Around the Craft Demo cluster the ordinary families: Java, OOP and design; arrays, strings and hashing; trees and graphs; DP and matrix. The problems are familiar and the grading is not. Clean layering and a couple of unit tests that actually run move the write-up further than shaving a pass off your loop, and edge cases you raise yourself count for more than the ones the interviewer has to go looking for. That testing instinct is one Intuit builds into its own tooling: it open-sources the Karate test framework on GitHub.

Most of what Intuit asks in design is generic, which is worth knowing before you spend a week reading about tax engines. The prompts that recur are a URL shortener, an autocomplete, a photo store, a notification service, and above all a rate limiter you have to actually build and then make work across instances.

One prompt carries the company's fingerprints. The first 120 users get a free seat, and you have to guarantee exactly 120 while the requests arrive at once, which lands you in locking versus atomic decrement, idempotency keys, and isolation levels. Real-time metrics and leaderboards recur as well. The worked designs here cover both of those, plus a generic HLD playbook, a tax rules engine, and financial data aggregation. Those last two are color to bring if a round turns product-domain, not something to plan the week around.

Whatever the prompt, the grading leans on financial correctness. Money in integer cents, an operation that can be retried without charging anyone twice, an audit trail that reconciles, data you would not want landing in a log. An interviewer who hears you reach for those unprompted stops wondering whether you have shipped anything that touches money.

Intuit is public, so the equity is liquid stock you can price on any given morning, and that price is currently the whole conversation. The shares are down roughly two thirds from their July 2025 high while revenue kept growing, which means a grant sized in dollars buys far more of them than it would have a year ago, and it means anyone quoting you last year's package math is quoting fiction.

Bands are narrow across every component and Intuit is not known to go above them, so the realistic target is the top of your band rather than beating a competing offer. Base moves least. Equity carries the most leverage at mid and senior level, and the sign-on is the friendliest lever in the building, regularly won by candidates with nothing competing to wave. The annual bonus is a percentage of base fixed by your level, so you do not negotiate it directly, though every dollar of base quietly lifts it.

Average total compensation by level, per Levels.fyi (US, as of Aug 2026); self-reported and rounded. Intuit labels levels by title, with no numeric codes, and equity is public INTU stock. The overall median is about $260K. Principal (Architect) and Distinguished Principal sit above these on thin samples, roughly $570K to $1.1M.

Negotiate the components together, once the debrief has landed positive and the offer is still being written. Push on one bucket in isolation and the recruiter just moves money out of another. Ask what the top of the band looks like for your level and whether a sign-on is available at all before you build a counter, and value the shares at what they trade for on the day you sign.