J'ai postulé via un recruteur. Le processus a pris 2 semaines. J'ai passé un entretien chez Google (Mountain View, CA) en oct. 2017
Entretien
Connected to a recruiter and got onsite in MTV. Google's recruiter is friendly, set up the onsite. 5 rounds of tech interviews. Although the position title is machine learning engineer, the interviews were all coding. Really, no questions about machine learning. It is strange that, although you come up with your own algorithm, the interviewers did not like it and asked you to think and follow their algorithm. It seems you have to give their "standard correct" solution. Two interviewers intently stated the questions very very vaguely such that it is about understanding the question, not about algorithm design. No offer in a week.
Overall, Google looks for software engineers if you applied yourself. If you want to do machine learning in Google, you probably need to be famous or some famous people know you.
Questions d'entretien [1]
Question 1
Questions from LeetCode may be asked. But if they sense you knew the problems before, they will change the problems, which they may not be very familiar themselves.
Standard interview. I was contacted by the recruiter and was matched with a team pretty relevant to my job. One phone interview, one fit chat, 4 rounds of onsite interviews.
Questions d'entretien [1]
Question 1
Implement a MHA class, together with some modeling questions related to SOTA models.
pas eu d'entretien. jamais répondu. pourtant j'avais un super cv et une recommendation d'une personne travaillant la bas. donc très déçu de l'administration chez google. je pense que les cv passe par des IA qui prennent que des gens de Stanford
Na entrevista foram feiras perguntas condizentes com o cargo, sem fugir do tema, Estrutura de modelos e experiencias profissionais. Além disso, boas praticas com modelos de IA e ciclo de vida de modelos.