A governed analytics platform that translates business questions into KPI planning, semantic retrieval, SQL generation, Microsoft Fabric execution, validation, visualization, and decision-ready insights.
Impact
KPI retrieval, SQL execution, validation, and insight generation within one governed workflow
Hiring signal
Multi-agent orchestration, Python backend ownership, analytical reliability, Microsoft Fabric integration, and typed response design.
My role
Lead developer · FastAPI backend, agent orchestration, SQL generation, validation workflows, Redis state management, Fabric SQL integration, and response contracts
Status
An interactive trace is available.
The problem
Business users relied on static dashboards and ad hoc BI support for questions that fell outside predefined reports. Each new request required manual SQL development, report changes, or a separate analytical cycle before a reliable answer could be delivered.
My approach
The platform uses a FastAPI orchestration layer to coordinate intent classification, query decomposition, KPI planning, metadata retrieval, semantic enrichment, SQL generation, Microsoft Fabric SQL Endpoint execution, result validation, chart construction, and insight synthesis. Each stage exposes a defined contract, allowing failures to be isolated, retried, and reported transparently.
Architecture judgment
A single text-to-SQL prompt is too fragile for an enterprise KPI catalog. Separating planning, retrieval, execution, validation, and response generation creates explicit control points, safer retry boundaries, and a traceable analytical output.
Architecture
The complete execution path, service boundaries, data stores, security controls, validation stages, and operational telemetry.
Implementation
Technology
Interactive trace
Run the trace to see the main workflow stages.