I spent some time on core services within amazon. I think theres just no upside in major rewrites when the system is working. Any major outage with cloudfront is a huge liability for the company, and so AWS decides to just pay their engineers a ton of money to sit there and hammer out operational issues day by day. This is really the secret sauce of AWS, industry leading operations/stability/backward compatibility.

On some teams, the work is actually not SWE work, even though people think it is. Its really an on call rotation, with some product related work to keep people busy while they recover from on-call. Managers will never admit it though.

There will be good SWE work, but some guy who has been in that same org for 6+ years will quickly snatch the best parts. Less tenured engineers will slog