The process seems to have 3/4 phases: 1) initial CV screening, 2) exercise, 3) huge exercise, 4) interview in person. 1 and 2 were the industry standard. Number 3, however, was a complete insult to any professional. They want someone with specific experience in their set of ideas, or a lot of time to waste. The 3rd exercise was extremely long, with dedicated literature and statistical models, which obviously aren't part of repertoire of most data scientists. Between reading the models and completing the challenge, I spent around 10h (half understanding the models and all statistical assumptions); I decided not to spend more. Despite having an extremely long response, my application was rejected since I did not test every single statistical hypothesis in one of the questions (not that I didn't know how to do, I just thought no respectable professional would ask so much for a screening exercise). Bottom line, I'm glad I wasn't accepted because I certainly don't want to be working at a place where I'm supposed to know it all in advance, or work for free (which seems to be expected from some of the reviews).