Length variesFormat Call with a recruiterThey grade Why Waymo and a short project walkthroughHeads-up The recruiter's preview of the format is often inaccurate; prepare more broadly.
Length 45 minFormat Coding: a classic algorithm problem or an original, under-specified problem framed around self-drivingThey grade Clarifying the problem, your own test cases, iterating from brute force
Length 45 minFormat Usually two rounds; some give you the interviewer's buggy code to find and fix, some are read aloud with no skeletonThey grade Clarification, trade-offs between approaches, your own main and tests
Length 45 minFormat Design discussionHeads-up Prompts increasingly come from the interviewer's own team (fleet data collection, maps, simulation) rather than generic web scale.
Length 45 minFormat No coding: reason about metrics and experiment design for a driving or simulation scenarioThey grade Hypotheses, metrics, confounders and stratification by scenario; acknowledging the gap between simulation and the roadHeads-up Appears in about half of SWE and MLE onsites.
Length variesFormat Behavioral or hiring-manager conversationHeads-up Senior loops swap one behavioral slot for a project deep dive. Some pipelines start with a non-technical hiring-manager round that is a real gate.