Building technically usable and available Data Products within the Data Mesh is one thing, but making them understandable for both humans and AI and interoperable between domains is a whole another challenge. At the core of this challenge is semantics.

Semantics is the missing piece in many Mesh implementations. Governed business metadata (as opposed to purely technical metadata) is needed to build context around the data - and human users and LLMs alike are hungry for context.

In this talk, Juha Korpela, an expert in data modeling, metadata management, and Data Products, explores ways to capture and share the semantics of a Domain or a Data Product. We will discuss the differences between solution-level and domain-level semantics, and look for ways to cross the gap between ubiquitous domain languages in order to achieve true semantic interoperability. Data models, ontologies, knowledge graphs - it's all about context!