The first round of interviews is where they ask some coding questions on Python and let you write codes on a whiteboard, followed by some questions regarding the statistics testing and behavior questions. Their requirement of skills is actually different from what they described on their career page.
Interview questions [1]
Question 1
Write a simple algorithm to decide on when to buy and sell stocks to maximize profit
I applied through an employee referral. The process took 2 months. I interviewed at Amazon (Bellevue, WA) in Dec 2019
Interview
Two phone screens- Both entirely technical, and no behavioral questions asked at this point. Basic questions on Statistics, Machine Learning and Probability. Case questions were also given to me, they wanted to hear my thoughts on a Supply chain project at a higher level, what kind of data would I use, what hypothesis would I assume, and how would I test the validity of it. I was invited to onsite which consisted of six non-stop interviews with the hiring manger, the leadership team and fellow Data Scientists at the position I was interviewing for. The p[position is like a fancy Business Analyst position, you are only responsible for getting the data ready, and hand it over to a Research/Applied Scientist that works on the models. More behavioral questions than I had imagined. I expected about 50:50 after reading Glassdoor reviews but the ratio of behavioral to technical would be 80:20 if you ask me. They were all questions around their Leadership principles and make sure you don't repeat your answers to different interviewers. They expect you to use multiple situations.
I applied online. The process took 2 weeks. I interviewed at Amazon (Bengaluru) in Dec 2023
Interview
One telephone screening, Three Technical and behavioral interviews. One of them included coding questions where I was asked to type out the answers. The last round is hiring Manager Interview.
Interview questions [1]
Question 1
What are some of the challenges you facing in doing your work and how did you handle them?