In this talk, I present our ongoing work on utilising edge-computing to improve the scalability and privacy of user-centred analytics in the context of personal/IoT data. I present an architecture where devices and resources centred around the user, collectively referred to as the edge, can complement the cloud for providing privacy-aware, yet accurate and efficient analytics. I then present the evaluations of the proposed framework for applying privacy-preserving deep learning techniques on a number of exemplar applications, and discuss the broader implications of such approaches for future systems such as the Brave Browser, the Databox platform, and health monitoring applications.