Waymo Machine Learning Engineer Interview Guide

Waymo tests machine learning engineers at the coding depth most companies reserve for software engineers, and expects ML depth on top of it. The design round draws on Waymo's own driving data, with prompts like retrieving similar scenes from a large dataset or evaluating a perception model. Senior and staff candidates go through the same interviews across Waymo's major offices, and Waymo doesn't set your level until they end.

This guide breaks down each stage of the MLE interview process, what Waymo interviewers look for, and how to prepare with real example questions, actionable tips, and resources.

Waymo MLE interview process

Waymo's machine learning engineer interview process puts a 45-minute algorithmic coding screen before the final interviews, and most candidates who don't advance are cut there. A recruiter prep call sits between the stages, and it's where you find out what each final round expects.

As an Alphabet subsidiary, Waymo hires like Google: a hiring committee reviews your interview feedback before any offer, you interview without knowing your team, and Waymo uses Google's L3 through L7 level ladder.

Here's what the interview process can look like:

1. Recruiter screen: 30 to 45 minute call covering background, work authorization, interest areas, and the team's language requirements
2. Technical coding screen: 45-minute live coding round in CoderPad, centered on data structures and algorithms
3. Recruiter prep call: Short scheduling call where you can ask what each final round covers
4. Final interviews: 4 to 6 rounds of roughly 45 minutes each, covering coding, ML coding, ML system design, and 2 behavioral rounds

The interviews run 4 to 8 weeks from recruiter screen to decision. Team matching happens outside that schedule and can add weeks or months, and your interview feedback expires after roughly 12 months.

Waymo recruiter screen

The Waymo MLE recruiter screen is a short call covering your background, work authorization, and reasons for wanting the role. Recruiters also name the technical requirements for the specific team, including whether the position requires C++.

Expect questions about the machine learning projects you've led and how much complexity each one carried. Waymo recruiters move quickly through this call, so keep your background summary tight and specific.

This is the first of 2 calls where the recruiter gives you information that changes your prep. Write down the team name, the language requirement, and the round structure before the call ends.

Interviewers look for:

• ML project ownership: Whether you've led machine learning work with genuine technical complexity
• Language fit: Whether your Python and C++ experience matches what the team requires
• Work authorization: Whether you need visa sponsorship and can work from the listed location
• Motivation for Waymo: Specific reasons you want to work on autonomous driving and on this team
• Level signals: How much scope and cross-team coordination your recent work involved

Recently asked questions

Here are some real interview questions reported by candidates:

• Walk me through what you're doing now, what you've done before, and why you're interested in Waymo.
• Talk me through your work experience in the past year.
• What are your areas of interest within machine learning, and what's your availability?
• Would you require visa sponsorship?
• Why do you want to leave your current job?

Waymo technical coding screen

The Waymo MLE technical coding screen is a 45-minute live coding round on data structures and algorithms. Waymo engineers run it themselves in CoderPad and give you one harder challenge in place of several easier ones.

Waymo hasn't published a policy on AI coding assistants, so prepare to work without one and confirm the rules with your recruiter. Google began piloting an AI-assisted round in the second half of 2026 for junior and mid-level engineers, and that pilot leaves the algorithmic interviews unchanged.

Most questions test graph and tree traversal, and the setup is usually a concrete system like an org hierarchy or a road network. Your code has to run, and the interviewer may ask you to write tests against it or account for its runtime.

Interviewers look for:

• Working solutions: Whether you reach running code inside the time available
• Traversal fluency: How confidently you apply breadth-first and depth-first search to unfamiliar structures
• Complexity reasoning: Your ability to state the runtime and memory costs of your approach
• Independence: How far you progress without hints
• Query efficiency: Whether the data model you build supports fast lookups afterward

Recently asked questions

Here are real, recent interview questions reported by candidates:

• Given an organizational hierarchy, how would you model it so you can quickly find a person's direct reports and everyone under their department? How would you answer those queries efficiently once the hierarchy is built?
• Implement depth-first search to find the connected components of a disconnected road network.
• Implement K-means clustering from scratch.
• Find the lowest common ancestor of 2 nodes in a tree.
• Compute a 2D prefix sum across a grid.
• Design a structure that maintains average speeds per road segment and supports add, remove, and query.

Waymo recruiter prep call

The Waymo recruiter prep call schedules your final interviews and answers your questions about the rounds ahead. Waymo doesn't score this call, so ask which language each round expects, whether any round requires NumPy or PyTorch, and how many final interviews you'll have.

Waymo sends a written prep document beforehand that tells you to practice data structures and algorithms. That list is incomplete, and the coding rounds also ask for ML implementations like clustering algorithms.

Waymo applies the word "coding" to 2 different rounds. One covers algorithms and the other covers NumPy data manipulation and model implementation, and each needs separate preparation.

Waymo applied coding interview

The Waymo MLE applied coding interview is a 45-minute round where the interviewer gives you a diagram of a driving scenario and asks you to compute an outcome from it. Recent questions have covered 2 objects colliding, a car navigating a grid of circular obstacles, and a ball moving through a grid.

The prompt takes longer to parse than an algorithmic question because the interviewer describes physical objects and their positions instead of a data structure. Walk through your reading of the diagram out loud and have the interviewer confirm it before you write anything.

The underlying work is stack, tree, and graph traversal plus simple simulation. Identify which objects are nodes, which relationships are edges, and what state you need to track per step. Build that model first, then write your implementation against it.

Interviewers look for:

• Prompt comprehension: How quickly you extract the computational challenge from a visual description
• Time management: Whether you leave enough time to write code after clarifying the setup
• Algorithm selection: Your ability to map a driving scenario onto a standard traversal or simulation approach
• Clarifying questions: Whether your questions narrow the scope or restate the prompt
• Working code: Whether your implementation runs against the scenario described

Recently asked questions

Here are real, recent interview questions reported by candidates:

• Given this collision diagram, figure out which of the 2 objects is bigger and which direction the resulting mass will move, then code it.
• Given 2 obstacles that collide, determine the collision outcome from a provided diagram.
• Navigate a car through a 2D grid containing circular obstacles.
• Trace a ball through a grid and return its final position.
• Find the shortest distance from all buildings on a grid.

Waymo ML coding interview

The Waymo ML coding interview tests whether you can manipulate data and implement machine learning components in code. Expect NumPy array work on small datasets, usually under 100 rows.

Waymo keeps the datasets small and tests you on the transformation. You'll reshape, tokenize, and aggregate raw data into a format a model can consume.

Every team runs this round in Python, whatever language the algorithmic rounds require. Know how to compute means across multiple dimensions, handle arrays with mismatched shapes, and write a clustering or regression routine without a library.

PyTorch and JAX come up on teams that train large models, and the recruiter will tell you which one matters.

Write vectorized code and get it correct before you optimize it. Interviewers weigh a correct, readable implementation above a fast one, and they'll ask you to justify the tradeoffs you made along the way.

If you do applied ML work daily, this round will feel closer to your job than anything else in the loop. Prepare for it anyway, because the gap between this round and the algorithmic ones catches most candidates off guard.

Interviewers look for:

• Array fluency: How cleanly you handle reshaping, broadcasting, and aggregation in NumPy
• Algorithm implementation: Whether you can write a clustering or regression routine from scratch
• Data judgment: Which parts of a raw dataset you keep and which you discard
• Numerical correctness: Whether your transformations preserve the properties the model needs
• Pace: How much working code you produce inside 45 minutes

Recently asked questions

Here are real, recent interview questions reported by candidates:

• Given 3D trajectory data, manipulate and tokenize it into a usable format so you can predict the next step from the current trajectory.
• Sample points from within a bounding box using NumPy.
• Implement batch normalization in NumPy.
• Compute a mean across multiple dimensions of an array with mismatched shapes.
• Implement gradient descent for linear regression from scratch.
• Implement max pooling with argmax across a feature map.

Waymo ML system design interview

The Waymo ML system design interview asks you to design a retrieval, ranking, or evaluation system for driving data, using techniques you'd apply at any large ML organization. Simulation and infrastructure teams sometimes drop the driving scenario and give you a generic ranking or recommendation prompt instead.

The interviewer runs this round as a discussion and uses follow-up questions to decide what you cover next, so you'll take light notes and sketch occasionally without whiteboarding a full architecture.

Transformer architecture, model evaluation, and inference efficiency come up as design topics on teams building Waymo's foundation models. Be ready to reason about attention, calibration, and the cost of running a model onboard a vehicle.

Interviewers look for:

• Structure: Whether you open with a framework before proposing components
• Requirement setting: How you define latency, recall, and safety constraints for the scenario
• Representation choices: How you propose to embed and index driving scenes for retrieval
• Evaluation design: Which offline and online metrics you'd use to prove the system works
• Responsiveness: Whether you follow the interviewer's questions toward the areas they care about

Recently asked questions

Here are real, recent interview questions reported by candidates:

• Given a video scene, such as a rainy driving accident, design a system to retrieve similar scenes from a very large dataset. How would this help build a training dataset for conditions like rain?
• Design a system that decides whether the car should stop at a pedestrian crosswalk.
• Design an evaluation system for perception models, including the metric you'd optimize.

Waymo leadership behavioral interview

The Waymo MLE leadership behavioral interview focuses 45 minutes on a single project of your choosing, and the interviewer will push into its technical detail further than most companies do. You'll need to explain architecture decisions and tradeoffs yourself, at the same depth you'd expect in a technical review.

Choose a project that ran long enough to hit real obstacles and pulled in more than one team, even if it failed. The interviewer spends most of the round on the obstacles, so pick the project where you can name 3 specific ones and what each cost you.

Open with a short, high-level overview, then answer the interviewer's questions about team structure, the hardest technical challenges, and the measurable result.

Interviewers look for:

• Technical ownership: How well you understand the more complex parts of the work you led
• Scope: Whether the project involved multiple teams and ran long enough to test your judgment
• Decision quality: Why you chose your approach and what you'd do differently now
• Impact: What measurably changed because of the project
• Reflection: What you'd do differently on the decisions that went wrong

Recently asked questions

Here are real, recent interview questions reported by candidates:

• Pick an important project and walk me through why it mattered, how you approached it, how you built the team around it, what challenges you faced, and what impact it had.
• What were the hardest technical challenges in the project, and how did you resolve them?
• How did you structure the team and coordinate the work across groups?
• Tell me about an impactful project that you led.
• Tell me about a time when you led a project from start to finish.

Waymo collaboration behavioral interview

The Waymo MLE collaboration behavioral interview moves through several short examples covering conflict, influence, and mentorship, and the interviewer asks about the technical substance of each one. Have the architecture, the constraint, or the disagreement itself ready to explain for every story you bring.

Most of the round comes back to working across team boundaries, so prepare examples where your work depended on a team you couldn't direct. Build each story around the case you made to that team and the terms you offered them.

The interviewer may also pose a hypothetical, such as how you'd get another team to commit to your roadmap. Answer with what the other team gains from helping you, then explain how you'd check the request against their existing priorities.

Interviewers look for:

• Influence without authority: How you get another team to commit when you can't direct them
• Mutual benefit framing: Whether your approach gives the other team a reason to help
• Conflict handling: How you resolve technical and organizational disagreements before escalating
• Mentorship: What changed about how another engineer worked after you coached them
• Technical grounding: Whether you can explain the technical substance behind each interpersonal example

Recently asked questions

Here are real, recent interview questions reported by candidates:

• Tell me about a time you had a conflict or disagreement with another team and how you handled it. How did you influence them when you depended on their support?
• Tell me about a time there was a technical challenge and a disagreement with a partner team.
• Can you provide an example of how you manage conflict?
• Tell me about a time you mentored other engineers.
• Give me a quick introduction of yourself.

How to prepare for the Waymo MLE interview

1. Practice pattern-based coding prep first: Focus on tree and graph traversal, grid simulation, and prefix sums. These patterns cover both coding rounds and the screen that decides whether you reach them.
2. Ask which coding round is which: The recruiter will tell you whether a given round is algorithmic or NumPy-based. Split your practice accordingly: traversal and grid patterns for the algorithmic rounds, and array manipulation plus implementing an algorithm from scratch for the ML coding round.
3. Confirm your team's language on the recruiter call: Perception, planning, prediction, and evaluation teams run their algorithmic rounds in C++, and recruiters state the requirement on the first call. Practice traversal patterns in whichever language applies, and start early enough that syntax isn't slowing you down in the screen.
4. Prepare one project with technical and cross-team depth: The leadership round covers a single project in detail. Choose one that spanned multiple teams and carried genuine technical difficulty, even if it didn't succeed.
5. Read Waymo's foundation model post before the design round: Waymo's write-up on its foundation model explains how one model powers the Driver, the Simulator, and the Critic, and how the onboard validation layer checks the trajectories the model produces. Its February 2026 follow-up on the Waymo World Model covers generative simulation of rare driving scenes, which maps directly onto the retrieval and evaluation designs this round asks for.
6. Practice out loud: The ML system design round is conversational and gives you no whiteboard. Run mock interviews where you talk through a retrieval or evaluation design with a partner.
7. Ask about team matching early: Waymo matches candidates to teams after the interviews, and it hires in batches with a limited number of team slots open at a time. Ask the recruiter which teams are hiring now and whether any have already expressed interest in your profile.

About the Waymo MLE role

Waymo machine learning engineers build the models behind perception, prediction, planning, and simulation for the Waymo Driver. Teams range from onboard model development to the infrastructure that trains and serves those models at scale.

Waymo machine learning engineers typically work on:

• Training and evaluating models for perception, behavior prediction, and planning
• Building simulation and scenario generation systems that test the Driver before deployment
• Designing evaluation and metrics systems that measure model performance against safety requirements
• Scaling training and serving infrastructure across TPU and GPU fleets
• Working with multimodal sensor data from camera, lidar, and radar

Waymo MLE experience and education requirements

Waymo lists 2 to 4 years of machine learning experience for mid-level MLE roles and 5 or more years for senior, staff, and infrastructure positions. Python is required across every team, and C++ appears as a requirement on perception, planning, prediction, and evaluation teams.

Deep learning framework experience is expected, with PyTorch, JAX, and TensorFlow named most often. A bachelor's degree in computer science or a related field is the stated minimum, and modeling roles list a master's or PhD with conference publications as preferred.

Additional resources

• Machine Learning Engineer Interview Course
• ML System Design Course
• ML Coding Questions
• Engineering Behavioral Interview Course
• Waymo Interview Questions
• Waymo MLE Questions
• Waymo's Careers Page

FAQs about the Waymo MLE interview

How long is the Waymo MLE interview process?

The Waymo MLE interview process takes 4 to 8 weeks from recruiter screen to decision. Team matching runs as a separate stage and can add weeks or months, so plan for a total closer to 3 months and treat a longer wait as normal. Passing the interviews doesn't guarantee a match, and candidates who don't land on a team can wait until their feedback expires at roughly 12 months.

Does Waymo ask ML system design questions about autonomous driving?

Waymo frames most ML system design questions around driving data, including scene retrieval, perception evaluation, and behavior prediction. The underlying techniques come from standard ML system design practice, so autonomous driving experience isn't a requirement. Simulation and infrastructure teams sometimes use generic ranking or recommendation prompts with no driving component.

What programming languages does Waymo expect for MLE interviews?

Python is required for every Waymo MLE role, and C++ is required or strongly preferred on perception, planning, prediction, and evaluation teams. Some recruiters state the C++ requirement on the first call and run the coding rounds in C++. Confirm the language before you start preparing. The coding rounds follow the team's requirement.

Does Waymo use the same interview loop for senior and staff candidates?

Waymo runs the same interviews for senior and staff MLE candidates and settles leveling after the final round. A hiring committee reviews packets for higher levels, which can extend the timeline by a week or more. Interns and new grads go through a shorter, team-specific process centered on algorithmic coding.

How hard is the Waymo MLE interview?

The Waymo MLE interview is difficult, and the algorithmic coding rounds account for most of that difficulty. Machine learning specialists who haven't practiced data structures recently find those rounds harder than the ML coding round, which covers familiar NumPy work. Waymo interviewers give minimal hints, so each round measures what you produce on your own.

Waymo is also far smaller than its Alphabet parent, so clearing the technical bar puts you in competition for a limited number of open team slots. Strong candidates get held at team matching for months, and that outcome reflects headcount instead of performance.

How much does a Waymo Machine Learning Engineer make?

Here are the reported compensation ranges by level for Waymo Machine Learning Engineers, according to Levels.fyi:

• L4 (Machine Learning Engineer): ~$291K
• L5 (Senior Machine Learning Engineer): ~$494K

Waymo grants equity as Waymo Units, and those units don't trade like Alphabet stock. Waymo has discussed running internal buybacks that would let employees sell a portion back to the company, though it hasn't publicly confirmed one. Treat the equity portion of a Waymo offer as illiquid when you compare it against a public-company offer.