He wants to get into the
practical application of machine learning, not machine learning theory.
This is a common mistake people new to the field make. You can be very successful by learning how to use machine learning frameworks, and that doesn't require lots of probability theory, mathematical statistics, and optimization. Not that it hurts.
vayarajesh, start using Tensorflow, you'll reach the ability to reason about problems to which machine learning can be applied, and how to apply it, much more quickly than starting by starting at the root of the tree of knowledge. You can always learn as much math as you want in order to dig as deeply as you want, but first get a sense of what you're dealing with.
