https://drive.google.com/file/d/1FrPnqXsSJQhTxzRb-bsnVHP99nC82sbF/view?usp=share_link

At Netflix, dozens of data systems coexist: GraphQL APIs, Avro/Flink pipelines, Iceberg tables, OpenSearch indexes, etc. Each system defines its own schemas, its own understanding of what ""Movie"" or ""Talent"" means. When these definitions drift — and they always do — we pay for it in broken integrations, manual reconciliation, and engineering hours spent answering: ""Which system has the correct definition?""

UDA provides a single, authoritative place to define business concepts. When a team defines ""Movie"" in Upper, UDA's domain modeling language, that definition automatically projects to all systems: GraphQL schemas, Avro schemas, Iceberg tables, Java APIs, etc.

This talk dives into the foundations of Upper, our conceptual framework built on RDF. Upper is self-describing (it defines what a domain model is), self-referencing (it defines itself as a domain model), and self-governing (it validates itself against its own rules). This foundation enables federated governance: teams own their domains without a central authority.

Key takeaways:

Data fragmentation — How disconnected models lead to duplication and quality issues, and how shared semantics reconnect silos.
RDF as semantic layer — How Netflix uses RDF as the semantic layer that generates schemas across technologies.
Model Once, Represent Everywhere — How a concept defined once projects to GraphQL, Avro, SQL, and more.