Session: Reproducible ML Experimentation With `rubicon-ml`

Capital One has open-sourced `rubicon-ml`, a Python library for machine learning that captures and stores model training and execution information, like parameters and metrics, in a repeatable and searchable way. `rubicon-ml` ensures full reproducibility and audit-ability for both developers and stakeholders alike. It is built on top of and designed to play nicely with many well-known open source tools and libraries like git, Scikit-learn, Dask, Plotly, and more! Users can log model metadata to `rubicon-ml` using the storage they’ve already set up without any additional servers required. After that, they can leverage `rubicon-ml`’s dashboard and other visualizations to explore their logged model metadata. This session will show users how to seamlessly incorporate `rubicon-ml` into their existing ML workflows.