What's New with rubicon-ml? Capital One's Solution for Logging the ML Lifecycle

Explore Capital One's open-source solution for standardizing the machine learning lifecycle in this 44-minute conference talk from All Things Open 2023. Discover how rubicon-ml, a Python-based tool, leverages popular libraries like fsspec, intake, and dash to enable easy logging, visualization, and sharing of experiment metadata. Learn to incorporate rubicon-ml into existing machine learning workflows, seeing how it integrates with common libraries such as Scikit-learn for model training and Dask for distributed processing. Understand how rubicon-ml's schema can standardize metadata logging for open-source models from packages like Scikit-learn, XGBoost, and LightGBM. Suitable for machine learning practitioners of all levels, from developers to data scientists and model review officers, this talk demonstrates how rubicon-ml can help track model iteration across all stages of the machine learning lifecycle.