I don't think "learning deeper applications of algebraic geometry" has anything to do with "deep learning" in a machine learning context.

My best guess is that OP had a course about rather abstract homological algebra, which he only grokked after learning about applications in algebraic geometry, which were "deeper" in some sense.