This case study demonstrates how to build a machine learning model using data from the Collaborative Drug Discovery (CDD) Public vault and apply the model to score compounds in a private CDD vault. Specifically, it shows how to: 1. Select active compounds from AZ-ChEMBL data in the public vault to train a model. 2. Build a model using the selected active compounds. 3. Generate predictions for approved drugs in a private vault using the new model. 4. Export the model for use in other software or share it with collaborators. The goal is to illustrate how models can be developed in CDD and leveraged across projects and groups to help drug discovery