In finance, there are lots of traditional Greek-letter names. Using gamma/delta/theta is probably _more_ immediately comprehensible to your audience than, say, secondOrderUnderlyingSensitivity/underlyingSensitivity/timeDerivative. (Of course, in this case you don't get to choose which Greek letters you use.)
There are other contexts where particular Greek-letter names are traditional. Again, usually you don't get to choose which ones you use (at least, not if you want to get any familiarity advantage from using them).
When doing algebraic manipulations, sometimes it's convenient to give one set of variables Latin-letter names and another "parallel" set corresponding Greek-letter names. E.g., if you're doing plane geometry, maybe one point is (x,y) and another is (xi,eta). In software, this is probably a strictly worse strategy than calling them (x1,y1) and (x2,y2) or something of the sort.
julia> 2π
6.283185307179586\pi → π
That works for any programming language. I’ve used it occasionally for my own programs but use plain ASCII if I expect others to work on the code.