Why is type-level function application written in reverse order compared to value-level function application in OCaml?

In OCaml, function application syntax has the function first and argument second:

f 123

But for functions at the type-level, it's written in the opposite order:

let x : int option = None

If option is thought of a function at the type-level, then why is it not written like this?

option int

Is there a technical or perhaps historical reason?