KEy: A Knowledge Engineering Harness
Who: François Scharffe, Knowledge Graph Conference founder
When: Wednesday, July 8, 2026, at 3:30 pm
Where: Winslow Building - Conference Room - 1140
Knowledge engineering is a well-established practice involving the processes and techniques used to construct and maintain knowledge graphs. Building knowledge graphs is often resource-intensive, requiring large teams of data engineers, data scientists, ontology modeling specialists, and domain experts working over multiple years.
While the field has matured over the past two decades—with standardized modeling languages, robust ontology practices, and performant triple stores now widely available—the emergence of reliable large language models has created new opportunities for automating knowledge graph construction.
This presentation introduces KEy, a knowledge engineering harness designed to automate knowledge graph construction in enterprise environments. KEy supports new ingestion tasks end to end, from source analysis and ontology alignment to graph population and validation. It also assists with investigating issues, answering questions, and resolving inconsistencies within the graph.
A distinguishing feature of KEy is that it is self-built and self-extending: it analyzes the work it performs, identifies recurring patterns and bottlenecks, and proposes or implements improvements to its own workflows. In doing so, KEy aims to reduce the operational burden of knowledge engineering while making enterprise knowledge graphs faster to build, easier to maintain, and more adaptable over time.