Play 34
Edge AI Deployment
On-device inference — ONNX quantization, IoT Hub sync, offline-capable for disconnected environments.
Deploy AI models to edge devices with ONNX quantization, model compression, and offline inference capabilities. IoT Hub manages device fleet, synchronizes model updates, and collects telemetry. Supports disconnected and on-premise environments where cloud connectivity is intermittent or unavailable. Container Instances run inference containers, Azure Monitor tracks fleet health, and automatic rollback protects against bad model pushes.
Architecture Pattern
Edge AI: ONNX quantization, offline inference, fleet management, cloud sync
Azure Services
DevKit (.github Agentic OS)
- agent.md — root orchestrator with builder→reviewer→tuner handoffs
- 3 agents — Edge AI Builder (gpt-4o), Reviewer (gpt-4o-mini), Tuner (gpt-4o-mini)
- 3 skills — deploy (109 lines), evaluate (106 lines), tune (104 lines)
- 4 prompts — /deploy, /test, /review, /evaluate with agent routing
- .vscode/mcp.json — FrootAI MCP with IoT Hub + OpenAI key inputs + envFile
TuneKit (AI Config)
- config/edge.json — quantization level, model config, memory constraints
- config/sync.json — update schedule, rollback rules, fleet targeting
- config/guardrails.json — model validation, inference safety checks
- evaluation/ — inference accuracy, latency benchmarks
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.