First step is initial screening with the online tasks on probability, where you share and are allowed to record your screen and mic. it lasts 30min. there is a 4-5 questions on statistics, with multiple-choice answers, and the last one with free form, where you should write a final result (some percentage).
Questions d'entretien [1]
Question 1
in a given week the probability that some A produces 8 products is more than 10x higher than B producing 8 products. the average prob. of 7 day produced num. of products is 8.
q: what is the prob. that B produce at least 2 products in a week
q: given 2 days, what is a probability that total number of products produced in a period is exactly 2
Multiple hours of online assessments and AI interviews. The company does not value your time. The questions were generally easy, basic probabilities and logic, but very time consuming. A single error in all those hours you spent will probably disqualify you from even getting to a human
Questions d'entretien [1]
Question 1
Just basic probability and logic problems, not related to AI research or programming at all
Soon after your application, you are asked to submit documents (like academic transcripts for instance) on their platform and to take online tests. You have several days to start them.
Six coding and problem-solving take-home questions, followed by three interviews: behavioural, problem-solving, and technical.
The technical interview was a mix of talking about past projects, knowledge-based questions on neural networks, and a walkthrough of how I would go about building a specific real-world feature.
Finally a reference check, with all three references contacted.
They seem to prioritise examining how you think and assessing your future potential over having lots of past experience.
Questions d'entretien [1]
Question 1
Several problem-solving riddles, similar to things asked in quant interviews