How much math do you need to know to create and use Machine Learning algorithms?

Maybe it's a slight exaggeration, but I loved this comforting answer from a question in this fun ML quiz -

How much math do you need to know to create and use Machine Learning algorithms?
>> You need to understand the basics and have good intuition, but today, almost everything is an API

Predictive Analytics and Cognitive Computing are different areas of machine learning (supervised and unsupervised). Both would make an application more intelligent.

Once you have built a predictive model with R, you can consume it from your apps by having AzureML expose it as a REST service and also SQL Server can run it in the database

Within the Cognitive Services portfolio, the Emotion API recognizes human emotions on pictures with the aid of a well-trained Neural Network. Emotion API now can detect emotions in videos, as well as static images

Related:
Machine Learning Algorithm Cheat Sheet
Data Science for Beginners