Intro Deploying machine learning models to production can be painful. As apps look to adopt machine learning to replace procedural logic, it becomes increasingly necessary for developers to consume these artifacts without writing purpose built, custom code per machine learning model. One difficulty is that much of the real-time data streams…