> You cannot run a real-time AI system as a single ML pipeline. It consists of at least an offline training pipeline that is run when you train a new version of the model and an online inference pipeline that runs 24/7.<p>Yeah, this is a big assumption, and one that isn't always true. There are plenty of cases where you want to have data trained only on a sliding window of recent data, and so might train and infer at runtime. It's particularly common if you move stuff down to the edge.