There are some who work in reverse engineering, CPU interactions or static analysis but those are often more senior positions within a company, are more research focused or specifically marketed as such; my role as a pentester is focused on dynamic testing from a blackbox perspective. Sometimes we are lucky to have architecture diagrams, API docs, or source code but they only serve to benefit the test from an external perspective. I don't analyze the code and report vulnerabilities there, I report findings from a perspective of breaking the application in runtime; the code only makes that easier.
Anyone here wishing to break into security to "be a hacker" might find web app pentesting to be the most familiar for developers (it's not far from skills used for UAT, QA and debugging) and provides a pathway down the OSI model. There are companies that will take strongly motivated and technical people to train into pentesters, as the field is vastly understaffed and it's easier to train someone on your methodology from day 1. However, this normally starts as Web App (it's where the money and clients are) and one can move into other areas over time.
I'm more than happy to provide more details or resources to those interested. My knowledge is more in the attacker area, but its possible to start in either side and pivot into the other. Time, patience, and a willingness to learn.