I applied online. The process took 1 week. I interviewed at Google DeepMind in Aug 2018
Interview
Had a coding round followed by theoretical round. Interviewers were nice and friendly. Had a good time. Enjoyed the questions that were asked. Coding round was a bit long, otherwise fine.
The Google DeepMind SDE interview process includes an initial recruiter chat, a math and coding comprehension round, technical coding interviews, system design, research discussions, and behavioral interviews, typically spanning multiple rigorous stages
Interview questions [1]
Question 1
Given a large dataset, design an efficient algorithm to find patterns or anomalies, and explain your approach and trade-offs
I applied through a recruiter. I interviewed at Google DeepMind in Dec 2019
Interview
I had two rounds of interview, back to back with around 25-30 minutes gap between both. The first interview was of 45 minutes and the other one of 1 hour.
Asked to code one-one question in both the interview. Questions were of easy level. The first interviewer was not very interactive with much difficulty; he was able to stretch the interview to 45 minutes.
Questions that were asked to code were much easy. Other than this, some questions from OS were asked.
1) Check if the given matrix is Hankel or not. (Code)
2) What is Memory Leak?
3) What can be possible reasons behind the error A mission symbol?
4) What is Dynamic Linking
5) For a given date increment it with the given number and tell the new date. (Code)
6) Why DeepMind?
7) Internship discussion and some questions from Resume.
8) What is <bits/stdc++.h> in C++? Include all the required libraries separately without using this.
9) Difference between vector::empty and vector::size? For improving complexity, use empty over size when possible.
And, they didn't give me an offer because I was fresher. And, they were looking for someone more experienced.
Interview questions [2]
Question 1
Difference between vector::empty and vector::size? For improving complexity, use empty over size when possible.