The Quilt Data Registry for Teams provides data science teams secure, reproducible access to their datasets in Python. Quilt (https://github.com/quiltdata/quilt) is an open-source tool that can package and version datasets at scale in a virtual "data package" that can be imported directly into Python. This enables reproducibility in machine learning and data analysis. Data packages can be deployed to wherever the data are needed--on premises, in the public cloud, or to individual data scientists' laptops. The Quilt Data Registry for Teams gives team admins the ability to audit and control data access across the team.