joblib parallel_backend not providing any speed up as is running on a single cpu

Question

I have a model.predict function with 65536 rows of data which takes about 7 seconds to perform. I wanted to speed this up using parallel_backend from joblib:

this is