Preparing for a Google interview in 3 months

Most three month plans are a list of algorithm topics with dates attached. That plan prepares you for one round. Google's process runs a recruiter conversation, a role-specific screen, then a set of interviews scored against attributes it publishes: general cognitive ability, role-related knowledge, leadership, and what it calls Googleyness. Written feedback goes to a hiring committee that never met you.

Two consequences shape the calendar. First, the coding round is one input among several, so a plan that spends twelve weeks on it leaves the rest unrehearsed. Second, because the record is a write-up rather than a recording, everything you fail to say out loud is invisible to the committee. Both facts argue for spending time on delivery, not only on material.

Before week one, get the resume the recruiter will read into shape, since it sets the level you are considered for. The free resume builder covers that part. What each round measures is covered under interview types.

You have twelve weeks and a dozen conflicting study plans. This section allocates them. The shape is simple: build in month one, apply under time pressure in month two, rehearse the whole loop in month three.

• Month one, fundamentals plus one weak round: arrays, strings, hashing, trees, graphs, and complexity analysis, at a pace that lets you re-derive rather than recall. In parallel, pick the round you least want to do and start there, usually the behavioral one. Write six stories from real projects, each with a decision you owned and an outcome you can state.
• Month two, timed and spoken: move to timed problems in a shared editor, and solve every one out loud, including the parts you get wrong. Add system design if the level calls for it. The goal for the month is not a higher problem count, it is being able to explain a chosen approach and a rejected one.
• Month three, full loops: run end-to-end mock rounds with follow-up questions, review each one, and repeat the round that scored worst. Stop adding topics. Late-arriving material displaces the rehearsal that actually moves a score.

Running those loops solo, with follow-ups and a review afterwards, is what the mock interview page is for.

The failure is rarely a missing topic. It is a plan that is all input and no rehearsal, and it looks productive until the first mock round.

Consider a data engineer twelve weeks out from an onsite who works through a large problem set on schedule and books a first mock in week eleven. The solution arrives on time and the explanation does not, because ten weeks of silent practice trained the wrong half of the task. Moving that first mock to week three would have caught it while there was still time to fix it.

The other common failure is schedule drift. Recruiter dates move in both directions, so write the plan as three blocks you can compress or stretch rather than as dated weeks. If the loop lands a month early, month two collapses into two weeks and month three stays intact. Cutting the mock loops to save the topic list inverts the priorities.

Across these twelve weeks, the honest use of an assistant is practice. Mock rounds with follow-up questions, a review of what you said, and repetition on the round you avoid are the work that changes a committee write-up, and none of it involves anyone from Google being on the call.

For live rounds, SubcueAI ships a desktop app for macOS and Windows that captures system audio and your microphone behind a floating local overlay, and a browser extension whose Side Panel captures a meeting tab's audio, the interviewer's side only. No meeting bot joins the call, and nothing is injected into the meeting page. Setup is walked through on the tutorial page.

The scope limits are worth stating plainly before you plan around them. A proctored assessment, a recorded round, a shared screen, or a company-managed laptop puts any assistant out of reach, and Google's coding rounds commonly run in a shared editor. Build the plan so that your score comes from rehearsal, not from a tool that may not be available in the room.