The technical round consisted of 3 interviews (math, ML, computer science), each one hour online. The interviews were more like conversations to grasp what knowledge you have. I did not qualify to be invited to the team-specific interviews.
Questions d'entretien [1]
Question 1
Math: - Describe what convex function is? Why do we want convex functions? - What is gradient descent? What are second-order optimisation algorithms? How can the 2nd derivative be used in an optimisation algorithm? - Describe Newton's method? Where does it come from? How can it be adapted to find a minimum in a function? - What is the formula for the Taylor series? Why does it make sense? Why is there a 1/n! in the formula? - What is Bayes' Formula? Can you give an example? - What is an invertible matrix? What are the criteria that apply? What are the procedures to invert a Matrix? Can you describe the Gauss Process to invert a matrix?
J'ai postulé via un recruteur. J'ai passé un entretien chez Google DeepMind (Zurich)
Entretien
Describe your research papers and then a coding questions plus tehcnical and theoretical questions in the field of deep learning relevant to the role. THen questiosns for myself and scenario based and what I would do if i had a problem with one of my coworkers etc
J'ai postulé en ligne. J'ai passé un entretien chez Google DeepMind (Londres, Angleterre)
Entretien
two data structures and algorithms problems over virtual interview. Standard leetcode prolems at medium/hard level. Interviewers were not helpful. Asked difficult follow up questions. One spoke poor english and was difficult to understand.
J'ai postulé via un recruteur. J'ai passé un entretien chez Google DeepMind (Londres, Angleterre)
Entretien
Three stages, 1) Recruiter call, 2) 2 Coding Interviews + 1 ML Fundamentals interview, 3) ML Design + an interview with the Lead Researcher
I failed at the second stage, where I sadly didn't do well in one of the coding interviews. Both the coding interviews were Leetcode mediums, and the ML fundamentals one was what you would expect: optimization, regularization, loss functions, transformers, as well as practical training and inference questions.
Questions d'entretien [1]
Question 1
A string parsing question where you evaluate a mathematical expression.