Many teams are considering data synthesis tools and systems to address privacy and security policies and regulations. By producing differentially private synthetic database instances, teams can avoid designing separate mechanisms for multiple target applications. However, few differentially private data synthesis methods explicitly preserve the structure and correlations of the underlying data in a database instance. To address this limitation, Ge et al. present Kamino, a constraint-aware data synthesis system that generates a synthetic database instance with differential privacy and structure guarantees from an input database instance. The authors also demonstrate how Kamino, which leverages a probabilistic database framework, could support use cases like the development of ML classifiers and answering marginal queries.