This sort of rigid adherence to a particular paradigm isn't restricted to enterprise developers. I've seen this sort of thing in practice (or attempted) from colleagues in data pipelines for machine learning and other very much not-enterprise settings.
I find such religious devotion rules common among
scientists who have to code even though they are
not programmers by trade. Religious overcommenting
etc. Whatever rule they happened to read in their
FORTRAN-for-beginners book.
ML people are probably more knowledgeable, but their
job is also often more like science than traditional
software engineering.
That doesn't match my experience (as a former academic scientist who moved into software development). The governing rule for us was always "publish or perish." Programming was a means (exploring a problem) to an end (advancing knowledge and publishing the results). I, at least, never came across the jargon for testing, maintenance, etc. until I moved out of academia and into this industry.
Oh I agree, scientists are far less inundated by software "best practices" than professionals. It's just that those which they have heard of, they often stick to religiously.
Personally I prefer the results of these scientists to much of what I see from real software engineers. It might be wreck, but it is only 1000 lines of wreck -- and not a sophisticated wreck created with powerful modern tools.