DATA GOVERNANCE PLATFORM
Governance Engine
Federated data governance with AI-powered auto-remediation, exposed via MCP to AI agents. Built for a global measurement & analytics company (anonymised).
PROBLEM
A global measurement & analytics company needed data quality, policies and lineage governed consistently across many data domains — and a way for AI agents to consume only governed data.
APPROACH
A graph-as-source governance architecture (Neo4j). Validation rules defined as first-class objects with executor functions; a coverage tracker measures which issues are caught; an automated remediation layer proposes and applies fixes for deterministic cases.
MY ROLE
Defined requirements, led delivery as AI & Data Governance Consultant via Uszatki Ltd. Designed the rule taxonomy, coverage rubric, and remediation protocol.
OUTCOME
A federated governance framework (catalogue, rules, policies, lineage, data quality) with AI-powered auto-remediation, exposed via MCP to AI agents, plus a conversational governance assistant.
HONEST CAVEATS
The demo below uses a fully synthetic dataset and synthetic rules — no client names, client data, or client rules are included. The domain (fictional media measurement panel) illustrates the rule categories and executor pattern from the real build.
GOVERNANCE SIMULATOR
Synthetic dataset · 40 rows · 10 active rules
RULES (10 / 10 active)
100% rule coverage active
INJECT FAULTS
100% synthetic data and rules. Built for a global measurement & analytics company (anonymised).