Google Software Engineer Interview Guide

The standard interview loop for Google software engineers looks like this:

• Google Hiring Assessment: ~30 min online 
• Recruiter Screen: 30 min
• Technical Phone Screen: 1-2 x 45-60 min rounds
• Full Onsite/Virtual Loop (4-6 rounds, typically 5)
  Coding: 2-4 x 45 min
  Code Comprehension: 45/60 min (select US teams may face AI-Assisted Code Comprehension round)
  Systems Architecture: 45 min (experienced candidates)
  Googleyness & Leadership: 45 min
• Hiring Committee Review & Team Matching: weeks to months

Google doesn’t publish any official round names and their official prep guide organizes what you should focus on by skill. We’re using names that were in the invites our candidates received to help orient you. Your recruiter will confirm which ones will apply to you.

Google says this process takes about 6-8 weeks but candidates have more often reported that it took them 2-3 months to get an offer, especially because team matching can stretch the timelines by a lot. 

If you’re interviewing to be an ML engineer, please head over to our Google Machine Learning Engineer interview guide.

There's two segments to look at here, and they are both important to understand.

1. Skills expected of SWEs: 

• code and system health, 
• code comprehension, 
• data structures and algorithms, 
• programming, 
• software testing and debugging, 
• systems architecture, 
• AI fluency, and 
• AI product integration. 

A new graduate and even a staff candidate are going to face the same rubric, but the expectations and the weights on these expectations are going to shift depending on the role and even experience level. So you are going to have to study and prep against this exact list.

1. Evaluation attributes:

Instead of having certain evaluation dimensions for specific rounds, every round you face in the Google SWE interview will be evaluated across some, if not all of these attributes:

1. General Cognitive Ability (GCA) is measured against how well you break down problems you aren't familiar with and is a focus upon how you think, essentially. This is the most heavily weighted dimension.
2. Role-Related Knowledge is pretty self-explanatory and craft focused. This entails the 8 skills we just discussed.  
3. Leadership here for Google is "emergent leadership". You step up when the moment needs you, and you step back when it doesn't. So the proof of leadership has nothing to do with your title or designation.
4. Googleyness is a characteristic that they like to describe as a mix of intellectual humility, comfort with ambiguity, bias to action and collaborative instincts.

Before anyone at Google talks to you, you probably have to sit through the Google Hiring Assessment (GHA) which is a work style assessment that has about 80-100 questions you have to answer. Candidates report that it lasts about 30 minutes but that’s just how long they took, it’s not a hard cut-off or anything. The assessment is a mix of questions and based on what some Prepfully candidates have received, we’ve tried to group them into three types: 

• Personality statements: You’ll have to rate statements about your personality at work between strongly agree to strongly disagree. Expect statements like "I recover quickly from setbacks” or “I am able to succeed in ambiguous work situations”, etc. 
• Workplace scenarios: These are situation based like a friend at a competitor asking for confidential data. You’ll have to pick your most and least likely actions.  
• Forced choice: You’ll have to pick “most like me” and “least like me” out of 3 somewhat reasonable statements. This type of question has been reported the least.

Whatever mix of questions you get, they’re going to go over the same point/statement in a bunch of different ways to see how consistent your answer is even when the question is asked differently. They won’t test any technical stuff here. 

People don’t usually fail the GHA but if they do, it'll end their application and they’ll be asked to re-apply.

This is a 30-minute screen over the phone or on Meet with a recruiter. They’re going to do a background check and get a read on your level to set the expectations for the rounds to follow. 

The questions they’ll ask are fairly straightforward: 

• What about your experience do you think is most relevant to Google?
• What teams have you worked on? What have you built and what is the scale you've operated at?
• What’s your timeline like? Are you currently interviewing anywhere else?

They’ll most likely ask you about comp and what you’re expecting but you’re not obligated to name a number. 

The focus of this 45 minute screen is entirely DSA, which you already know how to prep for. 

You’ll be coding in a shared Google Doc so you won't have an IDE or compiler. This takes a bit of getting used to, especially because you’ll be writing syntax without autocomplete.

Questions typically asked in the technical phone screen:

• Find the longest path in a graph. As a follow-up, return the path itself instead of its length.
• Derive merge sort from first principles, then extend it to run across multiple workers.
• Parse a string containing a mathematical expression and return its value.

The suggested approach for this round is to begin with clarifying questions and nail the requirements. Talk through your approach, write it out and then optimize and dry-run a few test cases to find any bugs.

If your performance is “borderline” you’ll typically be asked to go through this screen a second time.

For the onsite loop, you’ll sit through 4-6 45 minute interviews. While onsite interviews have been conducted virtually for most of the recent past, in-person interviews are making a comeback. Some Prepfully candidates attended in-person rounds and mentioned that they were expected to whiteboard during technical interviews.

For each round, every interviewer will grade your performance on a scale from Strong No Hire to Strong Hire.

AI tools are absolutely not allowed during your interviews except, of course, for the AI code comprehension that’s being piloted for a few select US teams. More details in the Code Comprehension section. 

1. Coding

Expect 2-4 technical coding rounds with each round being 45 minutes long. You’ll typically solve 1 problem per round. 

Spend the first 20-25 minutes arriving at a working baseline solution. The interviewer will use the remaining time to ask targeted follow up questions and introduce new constraints to see how seamlessly you can incorporate them. 

Some senior candidates (L4+) have reported being handed a whole class to implement with the real algorithm buried in one method. If any of your coding rounds take this approach, remember that your interface design and how cleanly you've separated concerns are pretty important here.

Some recently reported coding questions:

• Given a list of file names, return the length of the longest prefix shared by any two of them.
• Return the sum of all contiguous subarrays in an integer array where adjacent elements differ by exactly 1.
• Given points with integer coordinates on a 2D plane, return the area of the largest axis-aligned rectangle that uses four of them as corners.
• Given a list of tasks each with a type, where same-type tasks must be at least N intervals apart, return the minimum number of intervals needed to run them all.

2. Code Comprehension or AI-Assisted Code Comprehension

As of 2026, the Code Comprehension round can go in one of two ways

1. Traditional format: 45 minutes
2. AI-Assisted Code Comprehension: 60 minutes (for select US teams only)

Code Comprehension is defined as your ability to read unfamiliar code, map out structural relationships, identify issues and then propose and implement clean changes.

In the AI-Assisted Code Comprehension Round, the work you’ll have to do is somewhat similar except, you’ll have Gemini as an assistant. 

You’ll get an existing multi-file repo with issues and your job is to find them by prompting Gemini in the chat window. Gemini can read the context and hand you text or code but it can't touch your files, so the debugging is entirely your responsibility. 

The interviewer will assess your AI fluency and output validation skills. 

AI fluency will show in how you structure prompts in a targeted manner so that solutions you get from the model are actionable. Don’t take on a passive stance and throw open-ended prompts like “find all the bugs” into the chat window. It’s a bad look for both parameters that are being assessed. 

Both formats are a read on how fast you can get productive in a codebase you don’t own while working through tech debt and legacy implementations. 

This round is a staple at L5 and above and doesn't apply to "candidates newer to the industry" according to Google.

It is a 45-minute design conversation that you’ll be expected to lead. The prompts you’ll face will  be vague on purpose.

This round is a test of whether you can design a system under given constraints, pulling knowledge, theory, experience, and judgment together as you go. Google officially lists interfaces, distributed systems, limitations, robustness, and trade-offs as topics the conversation can go into. Computing foundations are also listed under “architecture” so be prepared for the interviewer to move from interfaces and resource estimation into context switching or memory caching.

Prompts reported by candidates:

• Design a real-time collaborative editor.
• Design Google's database for web indexing.
• Design a globally distributed key-value store.

Prepfully coaches recommend this approach for the system design round:

Scope the problem first -> which parts are worth solving, and why those -> talk the system through before you draw anything -> tradeoffs, unprompted -> take the constraints they add mid-conversation -> then scale to Google traffic.

The rest of the format is somewhat standardised across the industry, so our System Design guide covers the rest of it with some extra tips.

In this 45 minute round, you’ll face 4-5 behavioral questions. These questions typically open with “tell me about a time when” or “describe a decision you made” with a few follow-ups that’ll focus on what you did next or what came of it.

Since you’ll be reaching from your experience and past, Google will focus on evaluating communication, time management, decision making, flexibility, initiative, leadership, organization, and problem solving.

Some candidates recently mentioned that their behavioral round had a technical design discussion. They had a retrospective conversation about stuff they’d already built and had to defend its architecture to the interviewer. 

In this round, you’ll face questions like:

• Tell me about a time you made a pretty significant technical decision based on incomplete data.
• Walk me through a situation where someone on your team disagreed with your technical approach. 
• Talk to me about a project where you realised half way through that your initial design was just not going to work.
• Have you ever received a tough piece of feedback from your colleagues? 

We suggest that you prepare a story bank of 8-10 stories so you’re ready with an answer for every theme that a behavioral question probes. 

When you sit down to make this story bank we suggest that you take the classic STAR framework and build it out into STAR+ (Situation, Task, Action, Result + the X-Factor: Learning) to structure your stories. The X-Factor at the end, “Learning”,  is your definitive chance to demonstrate intellectual humility, and helps you reflect on a personal takeaway on how that situation changed you for the better. 

Read more about how you can effectively convey “Googleyness” in your response.

The Hiring Committee will receive a packet that contains every interviewer’s rating of your performance from Strong Hire to Strong No Hire. This committee consists of 4-5 seniors pooled from unrelated teams and they’ll review and decide if you should move on to the Team Match phase.   

For a few lucky candidates, team match takes a week or so because a team was already waiting on someone. More commonly, it takes 4-6 weeks, sometimes even slipping into months.

Verified 2026 median total comp for Google SWEs are roughly:

• L3: $212,000
• L4: $311,000  
• L5: $418,000
• L6: $589,000.
• L7: $837,000

View the latest compensation for Google SWEs on Levels.fyi