Introduction

AI-enabled coding interviews are new, and they're spreading fast. Meta, Shopify, LinkedIn, Canva, Uber, and a growing list of top companies now sit you down with an AI assistant and a real codebase instead of a blank editor and a single algorithm question. The format moves faster than the advice written about it.

This guide is built from conversations with candidates who just went through these interviews and the interviewers who run and grade them. Candidates told us where they got stuck and what they wished they'd done differently. Interviewers told us what they grade, and how they can tell when someone is in control versus just prompting and hoping.

What makes this format different

AI-enabled problems are much larger than traditional ones. Instead of implementing a single function, you might be dropped into a multi-file codebase and asked to fix a bug, add a feature, and make it scale — all in one session. Traditional interviews produce 30–50 lines of code. AI-enabled ones produce several hundred lines across multiple files. You're not writing all of it by hand, but you need to understand all of it.

The skills that matter shift too. Algorithm memorization counts for less, and reading unfamiliar code quickly counts for a lot more. You're also running two conversations at once — one with the AI, one with the interviewer — and most candidates underestimate how unnatural that feels until they're in it.

What interviewers are evaluating

AI makes you feel productive even when you're in trouble. You paste the problem in, get 200 lines of code back, and feel like you're flying. But if you never planned, can't explain what came back, and accept everything without a second look, the code can look great while you quietly bomb the interview.

Interviewers are not grading clever prompts. They're watching whether you are still in control of the work. Across companies, four things stay consistent:

1. Problem-solving and approach

The fundamentals haven't changed. Interviewers want to know you can understand a problem, break it down, and prioritize correctly. AI adds a failure mode that doesn't exist in traditional interviews: because the model responds instantly with complete-looking code, it can hide the fact that you don't actually have a plan.

Think out loud during this phase. "I'm going to take a couple minutes to read through this and make sure I understand the requirements before I start" signals confidence and keeps the interviewer engaged.

2. Control over the AI

Just about every interviewer we talked to said the same thing: "We don't want the AI making decisions. We want to see you making decisions and using the AI to execute them." The approach is yours to set; the AI carries out the plan you've laid down. The moment it starts making architectural calls while you sit and watch, you've given away the exact thing the interview is there to assess.

One Rippling candidate got explicit feedback that they "relied too heavily on AI even though their initial approach was correct." They knew what to build but let the AI make the implementation decisions — and that distinction cost them. At Canva, interviewers pause after each generation and ask "what does this code do?" If you can't walk through the generated logic confidently, it tells them you're not really directing the work.

3. Verification habits

AI will introduce bugs. It makes wrong assumptions about your data model, misses edge cases, and produces code that looks right but subtly isn't. Candidates who accept AI output without reviewing it leave a bad impression, even when the code happens to work. The review catches mistakes — and it shows the interviewer you're in control.

Good verification means running the code after each meaningful change, reading through what the AI generated to confirm it matches what you asked for, and testing before you move on to the next phase.

4. Communication

You're managing two conversations at once, and the hard part is that the AI generates code faster than you can narrate it — creating a constant pull to keep prompting in silence. That's the instinct to fight. Going quiet for long stretches leaves the interviewer with nothing to evaluate.

The fix isn't complicated: say what you're about to do before you prompt, read the output out loud as it comes back, and call out anything that looks off.

How to use this guide

Less time than you think. Got an interview this week? Read the four evaluation areas above, skim the two formats, and do one full practice run so the two-conversation juggle isn't brand new on the day. Got more time? Work through the fundamentals in order and practice until directing the AI feels natural.

The Fundamentals section maps directly to the four evaluation areas: codebase orientation and planning cover approach, driving the AI covers control, verification and testing covers exactly what it sounds like, and communication covers the narration piece. Read in order, or jump straight to your weak spot.

Once you've covered the fundamentals, check the company-specific breakdowns — each covers the platform, the format specifics, and what that company's interviewers care about.

None of the engineers who passed were doing anything superhuman. They stayed in the driver's seat while the AI did the typing — and that's a skill you can build. The rest of this guide teaches it.