Went through the Intuit SWE hiring process earlier this year for a mid-level role in Mountain View (hybrid). I'll just document the coding portions because I could not find a clear breakdown before my loop and had to piece it together from old threads.
Online assessment (before phone screen): Two coding questions on HackerRank, 90-minute window. One was array/string manipulation, medium difficulty by LeetCode standards. The second was a graph/BFS problem, also medium. Nothing hard. It felt more like a filter than a real signal generator. Finish in 45 minutes and move on with your life.
Coding rounds during the onsite: Two separate 45-minute rounds with different interviewers, each with one problem. My experience: Round 1: Tree traversal with a twist. Not a textbook problem, but close enough that anyone who's done moderate LC prep would recognize the shape. I fumbled for a minute on edge cases and the interviewer was patient, gave a small nudge. Round 2: Design a simplified rate limiter, but coded. Not whiteboard, actually implement it in Python or Java. This surprised me. It was basically a coding + light design mashup. I used a sliding window counter and explained tradeoffs vs token bucket. That seemed to land.
Difficulty overall: I'd say medium. Not Google-hard. They're not trying to stump you with tree DP or segment trees. The bar felt like: can you problem-solve methodically, communicate, and write reasonably clean code. I left feeling like my thinking was evaluated, not whether I'd done this exact problem on LC before.
Languages I saw others mention: Python, Java, JavaScript all accepted. My rounds were in Python.
If you're coming from a bootcamp background or self-taught, this loop is genuinely approachable if you've done a few weeks of consistent LC practice at the medium level. You don't need to grind 200 problems.