We detail a new framework for privacy preserving deep learning and discuss its assets. The framework puts a premium on ownership and secure processing of dataand introduces a valuable representation based on chains of commands and tensors.This abstraction allows one to implement complex privacy preserving constructssuch as Federated Learning, Secure Multiparty Computation, and DifferentialPrivacy while still exposing a familiar deep learning API to the end-user. We reportearly results on the Boston Housing and Pima Indian Diabetes datasets. Whilethe privacy features apart from Differential Privacy do not impact the predictionaccuracy, the current implementation of the framework introduces a significantoverhead in performance, which will be addressed at a later stage of the develop-ment. We believe this work is an important milestone introducing the first reliable,general framework for privacy preserving deep learning.