AI Builder - Data Agents
THE OPPORTUNITY:
Build trusted conversational analytics and decision-support agents over governed enterprise data. The focus is semantic accuracy, traceable answers and business-relevant insight using BOTH Databricks Genie Agents and Microsoft Fabric data agents, including Fabric IQ Ontology.
WHAT YOU'LL DO:
Create and improve data agents that answer natural-language business questions using enterprise datasets and governed semantic layers. Configure Databricks Genie Agents over Unity Catalog datasets and metric views, including instructions, example SQL, business semantic expressions and verified answers. Build Microsoft Fabric data agents over Power BI semantic models, lakehouses and warehouses, including Prep for AI metadata and AI data schemas. Define entity types, properties, relationships, metrics and bindings in Fabric IQ Ontology; connect ontologies to Fabric data agents as semantic context. Investigate incorrect answers by inspecting agent run steps, generated SQL/DAX, model metadata and retrieval paths; create evaluation cases and accuracy benchmarks. Collaborate with data strategy, analytics and platform teams to establish data ownership, authorized access, freshness, quality, traceability and ongoing monitoring. Deliver insight capabilities for operational dashboards and decision workflows such as performance reporting, cloud cost analysis and vendor analytics.
REQUIRED EXPERIENCE:
Hands-on implementation experience with BOTH Databricks Genie Agents (formerly Genie Spaces) and Microsoft Fabric data agents. Practical ontology/semantic modeling experience, including Fabric IQ Ontology and entity relationships, metrics and data bindings. Strong SQL and DAX, with dimensional modeling and Power BI semantic-layer design fundamentals. Databricks knowledge of Unity Catalog, metric views, verified answers, tuning/benchmarking and the Genie Agents/Conversation API.Fabric expertise with semantic models, lakehouses or warehouses and Prep for AI settings that influence DAX generation. Ability to test agent accuracy, troubleshoot generated queries and provide verifiable, explainable responses from governed data. PREFERRED EXPERIENCEDeep Power BI semantic model development and Microsoft Purview governance. RAG over unstructured information using Azure AI Search. Exposure of Fabric/Databricks data agents through Copilot Studio, APIs or Azure AI Foundry. DELIVERY ENVIRONMENTWork in a dedicated, cross-functional AI delivery pod alongside business stakeholders, project management and technical partners. Deliver in fixed two-week waves against clear acceptance criteria; reuse shared patterns, complete architecture and security reviews, and work with independent QA/AI evaluation before release. Solutions move through governed development, UAT and production stages, with support and hypercare handoff. Responsible AI, data access, auditability and cost versus business value are built into the delivery process.