Key takeaways
Waymo's how-we-hire page sets out three rules on candidate AI use and never states an end-to-end timeline. A virtual onsite with "up to five interviewers" at "approximately 45 minutes" each, and three AI rules that give three different answers, published by a company that says it uses Gemini in its own daily work.
The AI section is the useful part, because Waymo drew the boundary in public and gave a reason for it. That tells a candidate what the interviews are measuring.
What are the stages of the Waymo interview process?
Waymo's how-we-hire page walks through four steps once you have applied.
Resume review and the first call. Waymo says its recruiting team reviews your resume once you have applied and reaches out if your skills and experience match, with a sourcer or a recruiter running that first conversation. Waymo then invites a question many candidates hesitate over: the first phone interview is "an excellent time to ask about the timeline and what to expect during the hiring process" for your specific role. The page itself never states a timeline, so the recruiter call is where you get one.
Phone or video interviews. Waymo describes "1-2 behavioral and/or technical phone or video interviews with a potential peer or manager," and tells candidates to be prepared for "behavioral, hypothetical, and case-based questions that cover your role-related knowledge." Waymo names a peer or a manager for this round, and the question types span three categories.
The virtual onsite. Candidates who advance are "invited to a virtual onsite," where they "meet with up to five interviewers, including potential teammates and cross-functional partners." Waymo expects "each of these meetings to last approximately 45 minutes."
The offer. Waymo makes it once it determines you are the most qualified candidate for the role, and the recruiting team handles compensation and onboarding from there.
Waymo does not total those figures up, so do the arithmetic yourself before you plan your week. At the top end its own numbers describe five 45-minute meetings in the final stage, a little under four hours of interviewing, and as many as eight conversations across the whole process once you count the recruiter call and the phone or video round at its published maximum.
One word in there is worth pausing on. Waymo calls the final stage a virtual onsite, so the interview loop runs remotely by default at a company whose product is physical. Treat that as a scheduling detail about the interviews alone. Waymo's early careers page describes a different arrangement for its intern programme, where "as a Waymo intern, you'll be hosted onsite at one of our locations and work in a hybrid modality." Neither that page nor how-we-hire states a work-location policy for full-time roles, so ask your recruiter.
Can you use AI in a Waymo interview?
Waymo answers this in three rules, and they do not give the same answer, so read all three before your first screen.
On coding and technical assessments, Waymo asks candidates to work independently: "unless specifically noted in the instructions, the use of AI tools or LLMs is not permitted." The conditional matters. Read the instructions on the assessment itself, because that is where any exception would live.
On live interviews, the rule is flat. "All live interviews (technical, behavioral, and design) must be conducted without the assistance of AI." All three named formats, no exception attached.
On application materials, Waymo leaves room. You "may use AI to help polish the grammar or clarity of your resume," provided the application still reflects "your own original work and authentic professional history." Polish is allowed. Authorship stays yours.
Read together, the three rules draw a line around evaluation. Where Waymo is measuring you, the work has to be yours. Where you are packaging work you already did, a grammar pass is fine.
Why does Waymo ban AI in live interviews while using it internally?
Waymo states the reason on the same page, and it is worth quoting in full: "while we use Gemini and other AI tools in our daily work to build the World's Most Trusted Driver, our priority during the hiring process is to get to know you." Waymo closes its AI guidelines on a summary line: "we believe in the power of human ingenuity."
There is a measurement problem behind that sentence, and vendor data shows its scale. CodeSignal, whose platform runs technical assessments for employers, reported in a February 2026 release that cheating and fraud attempt rates on proctored assessments "more than doubled, rising from 16 percent in 2024 to 35 percent" the following year. Those are attempt rates on proctored assessments detected on CodeSignal's own platform, a narrower claim than the share of candidates who cheat, and they are enough to explain why a company would write its policy down.
Waymo gives its own reason for the live-interview rule in the sentence right after it: "we want to see how you think, adapt, and collaborate in real-time." Our reading is that the three rules describe what the interviews measure. Where AI is barred, Waymo is scoring what you can do unassisted, in real time, while someone watches, which is a different preparation target from the one most candidates train for. Four-Leaf has written about where the AI interview copilot market draws its own line, and about the difference between real preparation and real-time assistance; Waymo has now put its version of that boundary on its careers page.
What does Waymo tell candidates to do in its interviews?
Waymo publishes five interview tips for its process as a whole, and they read like a scoring rubric turned inside out.
Ask clarifying questions, because "many interview questions you'll come across may be deliberately general" and Waymo says interviewers are watching for candidates to ask them before attempting a solution. The vagueness is deliberate. Define and frame the problem before solving it, and break a large one into smaller pieces to show a structured decision-making plan. Show your thinking as you go, because Waymo says "we're not looking for perfect answers, instead, we really want to understand your thought process and how you use data to inform decisions." Waymo also tells candidates to refine a first answer, though it scopes that one: "particularly in interviews for engineering roles, the first answer that comes to mind may need some refining."
Read together, those tips point at a narrated, question-first answer. Waymo does not state what any individual round scores, so treat that as our reading of its published advice. A candidate who hears an open-ended prompt, asks what is actually being optimized, names the constraint they are assuming, then improves their own first pass out loud is following Waymo's published instructions closely.
Those habits are also the ones an unassisted interview exposes. Asking a good clarifying question and revising your reasoning while someone watches is hard to fake and gets worse without rehearsal, and Waymo's live rounds carry no AI assistance. Four-Leaf's voice mock interviews exist for that gap between knowing your answer and delivering it under pressure.
What examples should you bring to a Waymo interview?
Waymo names four qualities by hand, which is enough to build a story set around. Its page tells candidates to "find examples you've had in dealing with ambiguity, complexity, prioritization, and gaining alignment across a matrixed organization."
Three of those are about the work. The fourth is about the org chart, and it is the one most candidates arrive unprepared for. Gaining alignment across a matrixed organization means getting a decision made among teams that did not report to you and did not share your priorities. If every story you have rehearsed ends with you shipping something impressive alone, half of Waymo's stated rubric has nothing to score.
The ambiguity item is specific too. Waymo's why-Waymo page describes the technical problems its teams meet as unique and highly ambiguous, which suggests interviewers want evidence you can operate without a specification. Choose the project where nobody knew what correct looked like when you started. Our guide to telling interview stories that land covers how to keep the ambiguity visible.
What is overrated
Memorizing a round count from a prep site. Waymo's own page gives ranges at both stages, one to two interviews on the phone and up to five interviewers at the onsite, and neither converts into the fixed sequence aggregator pages like to print. Prepare for the top of the range and treat the rest as upside.
Polishing answers until they are airtight is the other one. Waymo says in terms that it is not looking for perfect answers and that it wants the thought process, and a candidate who has over-rehearsed tends to deliver a finished paragraph with no visible reasoning in it. That is the opposite of what the page describes.
The playbook
- Ask your recruiter for the timeline and the role's specific loop on the first call. Waymo's page tells you to, which makes it an expected question.
- Read the instructions on any take-home or technical assessment before you start, because Waymo's AI rule there has an exception and the instructions are what trigger it.
- Rehearse out loud with no assistance, since that is the condition every live round is run under. Record yourself and listen back once.
- Build one story for each of Waymo's four named qualities, and make sure the alignment-across-teams one is a real cross-team example, not a solo project stretched to fit.
- Practice starting answers with a clarifying question and narrating a revision. Both are written into Waymo's published tips, and both feel unnatural the first few times.
- Block out close to four hours for the virtual onsite, and treat it as five separate 45-minute conversations with different people.
Where this is heading
More companies will publish an AI-use policy for candidates, and the ones that write it carefully will say something real about what they are measuring. Waymo's version is specific in a useful way: barred where it scores you, permitted where you are packaging work you already did, with a stated reason that admits the company uses AI itself.
For candidates that trend is clarifying. The interview is turning into the place where you demonstrate the part of your ability no tool can stand in for, and companies are increasingly willing to say so on the careers page. Waymo ends its AI guidelines on human ingenuity. Read that line as a description of the scoring criteria.