Katana Skipper is a machine learning workflow engine that orchestrates events across microservices to create executable ML pipelines. It provides generic engine and communication components that can be configured with any microservices, along with sample ML services demonstrating data fetching, TensorFlow model training, and prediction serving. The engine supports Docker, Docker Compose, and Kubernetes deployment, and can handle services in Python, JavaScript, or other languages.