Experienced CV engineers, how deep into the fundamentals do you go (mathematics, classical methods, creating algorithms from scratch)? Especially when so many open source tools do the math for you?

I've worked in CV positions ever since graduating university a few years ago. I come from an STEM background, so I've taken classes in Calc 1-3, linear algebra, Diff Eq., the works. My whole undergrad was applying math.

However, when I got into the field, I didn't have to do the math because opencv, pytorch, numpy, and other libraries like it had done it for me. I never had to develop a classical method from scratch, never had to do …

algebra algorithms computervision diff engineers fundamentals graduating linear linear algebra math mathematics open source stem tools university