ABOUT GRID42

Regulatory judgment infrastructure.

GRID42 is not a training platform. Not a compliance tool. Not a credentialing app. It is the infrastructure layer that makes professional judgment — the kind that survives regulatory scrutiny, executive review, and AI agent verification — legible, provable, and portable.

THE REINFORCING LOOP

BKT mastery below 0.60 triggers HAS_GAP in Neo4j. The Scaffold reads HAS_GAP, surfaces diagnostic, issues credential.

Course Factory subscribes to HAS_GAP events. Routes targeted cubelet content. Learner demonstrates mastery via APPLY-face assessment.

HAS_MASTERED written to graph. Scaffold reads it as prior on next diagnostic. Precision improves. Content targeting sharpens. Both compound.

INFRASTRUCTURE STACK

MCP Server Cloudflare Workers OAuth 2.1 + PKCE, Stripe metering, 94 tools

Knowledge Substrate Neo4j on Railway Unified vector + graph traversal

Object Storage Cloudflare R2 Flat-file canonical substrate — JSON cubelets, assets

Assessment Cockpit Vercel Subscriber management, assessment dashboard

Production Agents Claude Code Agent SDK 5-agent pipeline on Modal burst compute

Observability Arize Phoenix LLM traces at agent boundary

Credentials W3C VC 2.0 + SD-JWT Issuer DID: did:web:euler.center

Ontology Pipeline SHACL + SKOS VOCABULARY → TAXONOMY → THESAURUS → ONTOLOGY.ttl

WHY REGTECH

The primary artifact is a judgment diagnostic result and verifiable credential. The learning is the mechanism, not the product.

RegTech market: $15B (2024) → $85B+ by 2030. Vanta raised at $1.7B. Drata at $2B. GRID42 sits alongside compliance infrastructure, not courseware.