Google Software Engineer Interview Process: What to Expect
Devana Team
Google says less than Amazon about the exact steps of its engineering interviews, but it has said a great deal about how it interviews, much of it from its own research into hiring. That is the more useful half: it tells you what the person across the table is doing, even when you do not know the question.
The process
Candidates commonly describe the same broad shape: a recruiter conversation, one or two technical screens, then a set of interviews, often four or five. Those mix coding, system design for more senior roles, and a behavioural round usually called "Googleyness and Leadership". These details come from candidate reports rather than a Google page we could check, so expect variation by role and level.
The decision itself is not made by your interviewers alone. Laszlo Bock, Google's former head of People Operations, described in Work Rules! (2015) how hiring committees review the interview feedback and decide, which is why what your interviewer writes down matters as much as how the conversation felt.
How Google assesses
Structured interviews and rubrics
Google's own re:Work guide to structured interviewing describes the approach: the same vetted, role-relevant questions for every candidate in a role, comprehensive notes of what the candidate said, standardised rubrics so every reviewer shares a picture of a good, mediocre and poor answer, and trained, calibrated interviewers.
Four attributes
In Work Rules!, Bock names the four attributes Google assesses:
- โ General cognitive ability: how you learn and reason through a problem you have not seen, rather than what you have memorised.
- โ Role-related knowledge: whether you have the skills the job actually needs.
- โ Leadership: including stepping up when a team needs it, not only holding a title.
- โ Googleyness: Google's word for how you work with other people.
Put those together and the practical lesson is simple. Your interviewer is filling in a rubric from notes. Reasoning you said out loud can be written down and scored; a correct answer you never explained cannot.
What our Google question bank looks like
Devana's question bank holds 72 Google questions, written to reflect how Google interviews: its products, its published values and the round types candidates describe. They are our questions, not leaked ones. How they divide up says a lot about where to spend your preparation.
| Round type | Questions | Examples |
|---|---|---|
| Coding | 28 | Autocomplete from a stream of queries; Top-K documents under a latency budget; Merge sorted result shards |
| System design | 16 | Design Google Search autocomplete; Design the web crawler; Design YouTube upload and playback |
| Technical deep dives | 16 | Diagnose a tail latency regression; Every cache between the user and the database; Make a 10,000-row page smooth |
| Behavioural | 12 | A decision made with incomplete information; Disagreeing with someone more senior; Something you built that failed |
33 of the 72 are rated hard. The topics that come up most often are Scalability, Systems Design, APIs, Databases, Hash Map and Strings.
๐ฏ Practise these out loud
Create a free account, open the question bank and filter by Google. Any question can be practised as a 30-minute voice interview with an AI interviewer that talks, asks follow-ups and runs your code. The free plan includes three interviews a month. Create a free account
Our Google questions lean on the problems Google's products actually have: search, autocomplete, crawling, ranking and sharded data. The follow-up usually asks what changes at far larger scale, because that is the question Google's systems answer every day.
How to prepare
- โ Narrate for the rubric. State your approach before coding, give the complexity unprompted, and name the trade-off you are making. Each of those is something an interviewer can record. What you actually get asked about complexity is a good place to start.
- โ Expect the scale follow-up. Solve it, then be ready for "what if the input does not fit on one machine?" Practise sharding, streaming and approximate answers.
- โ Practise without running code. Candidates commonly report writing code in a plain shared editor. Make sure you can catch your own bugs by tracing an example by hand.
- โ Prepare leadership stories even if you have never managed. Leading a migration, unblocking a teammate or changing a team's direction all count.
- โ Drill graphs and search problems. They are the natural material for Google-style questions; graph algorithms in interviews covers the templates.
Frequently asked questions
Does Google still ask brainteasers?
Not by design. In 2013 Laszlo Bock told The New York Times that brainteasers are "a complete waste of time" that "don't predict anything" (as reported by ABC News), and Google's structured interviewing guidance centres on questions relevant to the role.
What is Googleyness?
One of the four attributes Google assesses: how you work with others. In practice it is probed through questions about collaboration, disagreement and handling ambiguity.
Who makes the hiring decision?
Bock describes hiring committees reviewing the interview feedback, rather than the interviewers deciding alone.
