Production Text Classifier for Posts and Emails

System Design · Disney · Medium

Design a machine learning system that assigns category labels to text, where the inputs are posts and emails. Your design should address how raw text is tokenized, which features are engineered from it, and how the model is chosen. It should also describe both sides of the system end to end: the offline path used to train the model, and the online path that serves predictions.

Loading the rest of the question…