Play 48
AI Model Governance
Centralized model lifecycle management with approval gates, A/B testing, and audit trails.
Centralized model lifecycle management — model registry, version control, approval gates, A/B testing, safety evaluation, and audit trails satisfying SOX, EU AI Act, and enterprise risk governance requirements. Uses Azure ML for model registry, Azure Policy for governance enforcement, and Cosmos DB for audit state. Full lineage tracking from training data to production deployment.
Architecture Pattern
Model registry: gated promotion pipeline, automated evaluation, policy-enforced governance
Azure Services
DevKit (.github Agentic OS)
- agent.md — root orchestrator with builder→reviewer→tuner handoffs
- 3 agents — Model Gov Builder (gpt-4o), Reviewer (gpt-4o-mini), Tuner (gpt-4o-mini)
- 3 skills — deploy (228 lines), evaluate (141 lines), tune (197 lines)
- 4 prompts — /deploy, /test, /review, /evaluate with agent routing
- .vscode/mcp.json — FrootAI MCP with ML workspace + OpenAI inputs + envFile
TuneKit (AI Config)
- config/openai.json — evaluation model config
- config/governance.json — approval gates, A/B rules, drift thresholds
- config/guardrails.json — policy enforcement, audit retention
- evaluation/eval.py — Gate accuracy >95%, Drift detection >85%
Tuning Parameters
Machine evidence
FrootAI evidence lifecycle
This is an internal evidence maturity label, not third-party certification, accreditation, legal compliance, or a production guarantee. Missing or expired evidence demotes automatically; catalog claims cannot promote a play.
This play currently has design evidence only. A runnable scenario, endpoint evaluation, and build receipts are the next contiguous gates.
Repo Intelligence
v1A no-clone, revision-pinned map for agents and humans. Observed evidence is separated from inferred flow so the output stays useful without pretending to be a full call graph.