Pros
- very smart coworkers, you raise your IQ by just chatting with them; - good pay level; - awesome level of work dedication; - countless learning opportunities, technical challenges that are tackled on a daily basis are hard and fairly non-trivial; - hard-working people all over the place means that everyone is focused on delivering, therefore standard communication overhead and bureaucracy that is typical for big companies is almost not present (feels pretty much like a startup). Not a place for slackers - they won't survive even a week. You end up contributing much more than you ever thought you are capable of, which is a good feeling and builds up a great confidence level in your skills (at the price of getting quite exhausted though). - a good thing is that if you criticize against the work environment, no one will condemn you about it (aka there is no internal cult about Amazon), but will rather likely agree with you provided that you give some ideas for improvement - these are always very welcome, and if you are persistent and dedicated enough you can definitely have an impact, make a change, and this will be much appreciated; - people generally respect you a lot, I've never felt underappreciated and under-acknowledged even for a moment.
Cons
- terrible work-life balance, working long hours is standard and although it is an 'unstated expectation' it is rather assumed that you have to do it; - indescribably bad on-call, in some teams there are cases of people who quit after first couple of on-call shifts; - some organizations are focused on revenue only, and ignore dealing with technical debt. Tolerating this year after year leads to extremely bad code base which is impossible to improve, and is therefore just left to be patched manually during on-call shifts. In other words, the focus is put on curing the symptoms rather than fixing the real root causes. This is a classical example of organizations that simply do not learn from their mistakes and their past experience; - a key company value and tenet is to be extremely quick and fast with deliveries, which naturally comes at the cost of very poor to almost-zero quality. This is especially true for the case of terrible code bases, where pouring dozens of new features every month only exacerbates the problems and turns the on-call into a frantic clicking exercise, a manual maintenance of what is supposed to be an automatic service. The worst part is that next sprint you will be too busy to deliver even newer features, so you end up never taking care of fixing the already embarrassing v1 delivery from the previous sprint (that you thought to yourself "ok, it's not great now, but I'll definitely stabilize/polish it later" - actually this never happens). - this naturally leads to a huge attrition rate, people come & go like in a supermarket ... Just look at the thousands of vacant SDE positions that Amazon has at every moment. The average time spent in the company is so short, that it all looks like this is some sort of a temporary seasonal job, rather than a serious engineering activity that is part of a well planned long-term career and professional growth. Part of the reason why this still works is that for foreign workers it is a great opportunity to get their immigration status secured (aka PR, green card, etc.) if they can survive long enough. For others a major incentive is probably the good stock options package, but it requires that you survive really long.