Had a 2-session interview:
First was coding and CS reviews (algorithms, space + time complexity, asking about mutexes, semaphores).
Second was a "Statistics and Probability" and Code Comprehension interview which consisted of questions on binomial distributions and specifics on one type of neural network architectures (i.e. CNN's of which I had no experience).
Overall was tougher than it needed to be, and very specific. Found the first reviewer to be quite nice, and the second to be completely unprofessional and unfit to interview turning me off the company.
Questions d'entretien [1]
Question 1
Q: What is a mutex
Q: What does this code do (It was a convolution)
J'ai postulé via la recommandation d'un employé. J'ai passé un entretien chez Google DeepMind (Londres, Angleterre)
Entretien
~5 interviews: initial phone screen, technical screen, research interviews with 2 different interested teams, hiring manager/culture fit screen. Technical screen was the most intense, the rest were pretty laid back.
Questions d'entretien [1]
Question 1
What is linear independence, what is a pdf, how to invert a matrix efficiently, etc.
Don't apply to deepmind. They will waste your time by making you study for a technical quizz only to cancel all potential internship 2 weeks later. Deepmind is a disorganized company and this shows through lack of respect towards your time and energy.
The technical interview tests linear algebra, multivariable calculus fundamentals, probability and statistics fundamentals linear algebra, multivariable calculus fundamentals, probability & statistics fundamentals. It mostly covers first and second grade undergraduate contents.
