It is a complex topic, and imo the people who have done it have high variance in their outcomes. For example some brought their industry perspective back into a

It is a complex topic, and imo the people who have done it have high variance in their outcomes. For example some brought their industry perspective back into academic research and did well, and some just feel very ill-adjusted to academic research (I just cannot get over the fact that performance critical code gets written in often buggy Python in my area) and get homesick. I think some parts of CS academia just does not care about the software engineering (robustness, speed, etc.), and these can be tough to adjust to, so it varies by area too. Disclaimer: never been in industry, but knows several people (PhD students, professors) who were in industry once