Google Interview Questions — Page 211

Google·Machine Learning Engineer·Onsite - Coding / Algorithms

Jul 2026

Google ML Engineer interview, technical round focused entirely on implementing a Transformer block from scratch. Pretty deep dive, they wanted working code plus a real understanding of the math behind it.

Implement a multi-head self-attention module in Python (NumPy or PyTorch), taking an input of shape (batch_size, sequence_length, hidden_dim), splitting into multiple heads, computing scaled dot-product attention with optional masking, and projecting the concatenated output.
Implement a SwiGLU feed-forward layer where the gate and value projections run in parallel, apply Swish activation to the gate path, multiply elementwise, then project the result.
Put the attention and feed-forward components together into a full Transformer block with residual connections and layer normalization, returning output with the same shape as the input.
Explain the time and memory complexity of multi-head attention, and walk through how masking works across the attention heads.

“The shape tracking is where things get slippery.”

View Post

Google·Software Engineer·Onsite - Multi Round

Jul 2026

Prepped for a month, cleared the phone screen, got the onsite scheduled, and then got rejected standing in the lobby before a single interview happened. No feedback, no explanation, just a form rejection two days after they told me to go home.

What are the most likely reasons an onsite gets cancelled and then turns into a rejection with no interview ever happening?
Is there any value in asking for feedback when the rejection came before any real interviews took place?
Does a standard reapplication cooldown period still apply if the onsite was cancelled before it started?
Has anyone been re-engaged by the same team or a different team after something like this happened to them?

“Req closed, headcount pulled, internal candidate who got the nod before I showed up.”

View Post

Google·Data Scientist·Technical Phone Screen

Jul 2026

Google data scientist technical screen, one coding question focused on Monte Carlo simulation for hypothesis testing. Pretty niche for a DS role but made sense given the team.

Write code that simulates n coin flips many times, tracks the head counts across simulations, and uses that empirical distribution to compute a two-sided p-value for an observed number of heads. No closed-form stats formulas allowed.

“My first instinct was to reach for scipy.stats and they immediately said no.”

View Post

Google·Software Engineer·Onsite - Multi Round

Jul 2026

Went through the full Google software engineer onsite loop, four rounds total, and now I'm sitting in the waiting room refreshing my inbox. Felt pretty solid through the first three rounds but the last one was a mess of brute force and dry runs with no actual code on the board.

Implement a solution using a trie data structure (medium difficulty).
How would you handle various hypothetical workplace situations? (Behavioral / Googleyness round)
Solve an array and binary search problem (medium difficulty).
Solve a graph or matrix traversal problem (medium difficulty, given as a follow-up after finishing the first problem early).
Solve a hard graph problem involving a modified shortest-path algorithm (not the standard textbook version).

“Didn't get it to fully run, which still bothers me.”

View Post

Google·Software Engineer·Technical Phone Screen

Jul 2026

Google SWE interview with a classic caching problem. Nothing too exotic but the O(1) constraint is where people trip up if they haven't seen it before.

Design a fixed-capacity key-value cache that supports get and put operations, evicting the least recently used entry when full. Both operations must run in O(1) average time.

“The get and put parts are easy enough to describe but the O(1) eviction is where you have to actually think.”

View Post

Google·Machine Learning Engineer·Onsite - System Design / Architecture

Jul 2026

Machine learning system design round at Google for an MLE role. The whole thing was one big open-ended design question about building a real-time recommendation system, and they expected you to drive the conversation across a lot of different dimensions.

Design a real-time recommendation system for a large-scale mobile app that serves personalized results to millions of daily active users with sub-100ms P95 latency.
How would you design the feature store to ensure consistency between offline training and online inference?
Walk me through how you'd handle cold start for both new users and new items.
Given a strict latency budget, where would you use caching in the serving path and what TTL and consistency strategies would you apply?
What offline and online metrics would you track, and how would you set up experiments to evaluate the recommender?
How would you design the system to degrade gracefully if the model server is slow or unavailable?
Propose a baseline model and then a more advanced ranking model, and explain why you'd structure it as candidate generation followed by ranking.

“This was the entire interview, basically.”

View Post

Google·Software Engineer·Technical Phone Screen

Jul 2026

Google SWE interview with a string manipulation problem that looks deceptively simple but has a bunch of subtle invariants you need to nail down. The focus was heavily on proving correctness and complexity rather than just getting a working solution.

You have two equal-length strings made up of 'L', 'R', '.', and 'X' characters. 'L' tokens can slide left over empty slots but not right, not through walls, not through other tokens. 'R' tokens do the opposite. Can you transform the start string into the target string using valid moves? Design an O(n) time, O(1) space algorithm, explain your invariants, argue correctness, and walk through edge cases like mismatched walls, tokens separated by walls, and consecutive tokens with no room to move.

“I spent way too long trying to think of something clever before realizing the key insight is pretty mechanical once you see it: strip out the dots and the relative order of L/R/X tokens must match exactly between start and target, and walls must be in identical positions.”

View Post

Google·Software Engineer·Onsite - Coding / Algorithms

Jul 2026

Google SWE coding round, one problem the whole time. Tree validation, which sounds straightforward until you're actually in the room trying to remember every edge case you've ever seen.

Given an integer array where each element represents the parent of that index's node (with -1 marking the root), determine whether the array describes a valid rooted tree.

“I jumped straight to cycle detection and forgot to check for multiple roots first.”

View Post

Google·Software Engineer·Onsite - Coding / Algorithms

Jul 2026

Prepping for a Google onsite for a software engineering role and trying to figure out what the actual coding environment looks like on the day. Mainly wanted to know if it's a real IDE with autocomplete or the same barebones Google Docs-style editor from the earlier rounds, and whether they ask you to run code against test cases.

For the in-person onsite coding rounds, is the coding environment a proper IDE with autocomplete, or the same plain text editor used in earlier remote rounds? Do they ask you to run your code against test cases?

“This was my main anxiety going in.”

View Post

Google·Software Engineer·Onsite - Multi Round

Jul 2026

Went through the full Google Software Engineer pipeline over about two months, starting with a recruiter call and online assessments, then two online interviews, and finally two back-to-back in-person technical rounds. The process was long but Google covered all travel. Still waiting to hear back on the last two rounds.

Solve a HackerRank coding assessment with two problems of varying difficulty.
A medium-difficulty algorithmic coding problem with syntax-level precision expected, done on a shared spreadsheet.
Behavioral and situational interview covering how you handle workplace scenarios and team dynamics.
A hard dynamic programming problem involving path traversal, solved in-person.
A binary tree problem that expanded into five follow-up variations, each requiring meaningful changes to the previous solution.

“Fully solved the harder one, had some test case failures on the easier one which was embarrassing in hindsight.”

View Post