Data science is a research-heavy discipline concerned with staying on the cusp of technological advances. Scientists rapidly experiment with cutting-edge techniques and approaches, often coding in highly interactive, but highly unstructured, mediums like Jupyter notebooks.

How can a team integrate this experimental scripting and model handoff into a stable, defendable product?

International speaker with professional experience in the advertising and utility industries.

In my professional work, I focus on the stability and defendability of machine learning models. In my