In this video, we take data collaboration a step further by sharing a Generator (a generative model with metadate) instead of synthetic data in a Databricks Clean Room. Learn how to share a pre-trained, privacy-preserving synthetic data Generator that enables secure and flexible synthetic data creation. Using Databricks Clean Rooms, a collaborator can generate synthetic data on-demand, using a Generator that encapsulates the training of a GenAI model on original data.
We walk you through the entire process, from training the model and saving it in to UC, to uploading it into the Clean Room and sharing it securely. No data is shared — only the Generator, which allows the collaborator to create new, privacy-safe synthetic datasets tailored to their needs.