Python: Vectorized nonlinear least-squares optimization?

Question

Here is the basic scenario. Assume I have a vector of samples y (of length D) and a function f which returns another vector of length D called