Machine Learning Scientist applicants have rated the interview process at Amazon with 3.1 out of 5 (where 5 is the highest level of difficulty) and assessed their interview experience as 59% positive. To compare, the company-average is 57.5% positive. This is according to Glassdoor user ratings.
Candidates applying for Machine Learning Scientist roles take an average of 28 days to get hired, when considering 17 user submitted interviews for this role. To compare, the hiring process at Amazon overall takes an average of 28 days.
Common stages of the interview process at Amazon as a Machine Learning Scientist according to 17 Glassdoor interviews include:
Phone interview: 44%
Presentation: 13%
Skills test: 9%
One on one interview: 9%
IQ intelligence test: 6%
Personality test: 6%
Group panel interview: 6%
Other: 3%
Background check: 3%
Here are the most commonly searched roles for interview reports -
I applied through college or university. I interviewed at Amazon
Interview
I had two phone interviews, each lasted for about an hour. What I have seen Amazon do differently than other companies (and I like that) is that they go deep, during the interview, into the projects you mention in your CV. So, technical questions pop-up as you describe one of your projects. For example, You: I used this objective function and method X to minimize it. Interviewer: ok, so why is method X well suited for this, did you try Y? I know that implementing X has the following issues, and what are your assumptions for the model You: The model is blah blah because blah blah and this is non-convex etc etc, and we try to come around this by doing .....we test it like....(they ask a lot on how to test things)
There was also a programming part (machine learning related problem) and SQL. Can ask you something a bit more specific depending what you put on your CV or the role (Spark, GPU, or whatever you roll with).
Interview questions [1]
Question 1
I won't give details about the question as I respect the confidentiality of the interview. However, to give a general feeling, I think it doesn't hurt to mention the following. For example, code a class that implements a very popular ML algorithm. Even if the algorithm is very simple there are lots of possible improvements and generalisations, how to make it robust, efficient etc. Same thing for a class storing common data formats: dataframe, time-series, etc... how would you efficiently code access methods and/or storing according to the features of these data types?
Call with a recruiter followed by technical interview. The technical interview covered broad Machine learning questions. Then the coding part. The problem required prior knowledge of a calculus theorem, which in my opinion was not fair.
the interview process was nice and clear. Had two rounds of interview for the internship position. Both rounds were technical interviews. They asked my research experience and a case-study. Both had programming questions
I applied online. I interviewed at Amazon in Jan 2021
Interview
The interview process is very structured and clearly explained. The decision making on the employer's side was quite fast.
The only negative factor during the interview process was that no one showed up for my first scheduled interview.