Generative Programming for Constructing Efficient End-to-End Data Science Pipelines

Generative programming has shown great success in building large-scale systems with minimal engineering effort. In this talk, I will demonstrate how we can efficiently integrate such systems, particularly a data analytics system, and a machine learning system, to build efficient end-to-end data science pipelines. Specifically, we will show how such an integration can maintain best-of-breed performance for their specific workloads while eliminating performance bottlenecks in existing approaches to building such end-to-end data science pipelines.