They described their background. Then, I did the same. Then we turned to ML questions, which are listed below. There was a coding exercise, also listed below. Finally there was time for my questions.
Interview questions [6]
Question 1
What is the difference between bagging and boosting?
I applied through a recruiter. The process took 3 months. I interviewed at Amazon (New York, NY) in May 2020
Interview
Had 2 virtual on-sites, with 5 and 3 reps, set apart for 1 month. First set was more towards machine learning optimization, with some coding and a lot of leadership principles, second set was pure data science.
Did very well in the first set, but did not want to take that specific role, so they suggested to get a second set, which went very well as well. But after the interview recruiter emailed me that group decided to go with another candidate, and stonewalled all requests of getting on the phone. I wasted too much time for them and don't think such an attitude from a recruiter is professional, and i won't consider Amazon again just because of that.
Don't recommend to anyone, if you care about well-built recruiting system and reps professionalism.
I applied online. The process took 2 months. I interviewed at Amazon in Jan 2020
Interview
The process consists of two phone screen interviews of one hour each plus a on-site interview where the candidate has to pass multiple face-to-face interviews. If the candidate gets to the on-site interview stage, he/she will have to take a day-off (if currently employed) as the process takes 5 to 6 hours.
Interview questions [1]
Question 1
Phone screen interviews cover Data Science and Machine Learning depth and breath questions. The on-site interview focuses on Amazon leadership principles and coding questions. Furthermore, the candidate is asked to give an introductory presentation on machine learning.