A lot of attention is paid to the research and training of Machine Learning models, but very little attention is paid to deployment. In this session, we take a look at what it takes to get a model developed using tf.estimator or tf.keras into production.