Engineering knowledge base · MBSE RAG

The Ark knowledge base is a governable SysML/MBSE asset layer — sliced by methodology, diagram type, and task — that constrains AI4MBSE and future apps.

Planned as its own product — not merely a toggle inside AI4MBSE.

Not generic document RAG

Generic RAG chases semantic similarity. MBSE RAG chases modeling task fit: which MagicGrid domain, which OOSEM view, which Harmony-SE use case, which SysML diagram is active, and which project naming rules apply.

What gets ingested

  • Standards and internal norms
  • Methodology playbooks (MagicGrid, OOSEM, Harmony-SE)
  • Industry reference scenarios (see Cases)
  • Project conventions: naming, profiles, diagram gates

Preview scenarios

Used by AI4MBSE for conversational modeling and by future simulation products. Methodology · FAQ