In this talk I will outline a research project which aims to combine ideas from persistent homology and microlocal geometry for applications in stratification learning. Using stratification learning as motivation, we will trace through the historical context of this research with illustrative examples, touching on the theory of intersection homology developed by Goresky--MacPherson and the micolocal sheaf theory of Kashiwara--Schapira. I will conclude with a proposed theory of 'persistent microlocal geometry', and illustrate how this perspective can be used to study singularities in point cloud data.