The interview was similar to what is seen for most Data Science roles. There were 5 rounds in total. One test, 3 technical rounds and one non-technical discussion with their Data Science head who is based out of US. In each technical round they ask about your projects and then general questions around ML/stats/python. I cleared 4 rounds and then had the last round which was just around my interests and how I can contribute to the team. I felt it went pretty well. However, I just got a rejection mail post that. I requested HR to provide a feedback on what was the issue, however didn't get any response. This is so sick as after investing so much time and energy in an interview process you expect the organization to have decency in telling you what was missing. I hope they start showing a little more appreciation for the time and effort a candidate has put in their process by letting them know a proper feedback.

Applied online. First had a short phone conversation with a recruiter. I learned that most of the data science team are in India office. The Bay area data acience team doesn't have a manager yet! Next I had a 1-hour technical phone interview. The interviewer was a data scientist. She was very incompetent. She didn't know that ppv/sensitivity are the same as precision/recall!! There was a coding question in Python. Given an array of integers, reorder the elements such that non-zero element are in the beginning maintaining their orders followed by zeros. I was not supposed to use any other array in the code. I wrote a code that would look over elements and put any nonzeo element in the next available index then incrementing the index. Once all elements are visited, the code would fill all the elements after the last index by zeros. The interviewer was surprised and said she expected me to use two indices, one for non-zero and one for zero elements. It was possible but obviously an overkill.