Pros
Maybe it will look good on your resume. There are a lot of data and you may be able to do an interesting analysis (only if you are patient enough to wait for your query to run; see below). Work-life balance was not bad for me. There are a few smart and honest people.
Cons
Almost everything I wanted to say is in this review by another "research scientist II" on Aug 11, 2014 (Glassdoor does not allow a link in the review; filter reviews with "research scientist II"). Noise level is terrible because of overcrowded office with ridiculously noisy SDEs who have no notion that there may be somebody who needs a quiet environment for his work. We don't have access to academic journals and conference papers, and we need to ask people in academia to get a copy, which I believe is against the policy of publishers. 3yr vesting of 401k matching contributions was what I did not know before I left and what I regret the most. Amazon claims itself to be the world's most costumer-centric company. That in turn implies it does not care much about the employees. Basically, employees are dispensable and replaceable. Company is not investing enough for its internal infrastructures, because they are not directly connected to customer satisfaction. Databases have much less processing capacity than what is needed; the situation worsened much more since 2013 and it is not a surprise if it takes more than a few hours to run a very simple SQL query. I wonder how people, especially data scientist type of people, do their work with such a poor environment. Note that work environments for research scientists may be quite different depending on what area you are in. For machine learning and computer vision researchers, the company seems to be relaxing its strict policy on publications. However, on more traditional supply chain area that I was in, nothing seems to be changing; after all, we are probably not doing state-of-the-art research that is worth publishing. We are just coming up with ad-hoc patches for problems, without understanding anything and accumulating any knowledge. Promotions are rare and it looks tougher for research scientists compared to SDEs. While SDEs have clearer measurements on their work, it is often unclear how to measure the contribution of research scientists' work. Management, which is dominated by MBA-type of people, does not understand the value of long-term research and they are only looking at short-term cost savings, which are often estimated based on dubious calculations using inaccurate financial figures. There is no career path for research scientists. Internal transfers to other groups should be a good opportunity for employees to grow, but from my experiences, people are generally narrow-minded and reject you when you do not have past experiences in the exact area, even if you have good records of performance reviews. They basically treat you as just another external candidate. One last thing. My employment was terminated 2 days after I submitted a formal letter of resignation, even though I was requesting to leave one week later. Not only I lost 1 week of income, I also lost health insurance coverage one month earlier, because I would have been covered until the end of month if I quitted in the beginning of the month as I planned, but they have "fired" me in the end of previous month and the insurance was terminated at the same time. I didn't even have an exit interview. I couldn't believe how badly the company treats me after more than 2 years of service. Beware, because this can happen to a person like me, who was performing well enough to get promoted just several months ago. This whole thing illustrates how the company disregards its employees.