scipy curve_fit line is clearly wrong in highly skewed data

Question

I have a dataset that clearly demonstrates a curve-type pattern, but is not perfect. Here is the plain data.

My goal here is to fit a non-linear regression cu