In this talk we will first give an introduction and overview for one example, namely Topological Data Analysis, and discuss how it can be used to provide features for machine learning in a variety of settings. Secondly, we will discuss the ongoing challenges of translating this method into actual practice in these same settings. We hope this second half will give some insight into a basic challenge that faces all university research: it's much harder to apply your exciting new ideas than you think - most of the challenge is both how to prove its value and how to integrate it with what's already in use.