AI Platform Engineer

Summary

The AI Platform Engineer designs, builds, and operates the bank’s Microsoft Power Platform and Copilot Studio ecosystem, enabling secure and scalable AI‑driven business solutions across the enterprise. This role delivers production‑ready Copilot agents, custom Power Platform connectors, MCP servers, Dataverse data models, and reusable agent tools that extend Microsoft 365 Copilot to bank workflows. Through disciplined Power Platform ALM, Git Hub‑native delivery, and responsible AI practices, the engineer accelerates AI adoption while reducing risk and operational cost.

• Design, build, and deploy Microsoft Copilot Studio agents—including topics, knowledge sources, generative orchestration, authentication, and content moderation—that automate enterprise workflows across the bank
• Develop and maintain custom Power Platform connectors and MCP (Model Context Protocol) servers that securely expose internal banking systems and APIs to Copilot Studio agents and Power Platform makers
• Architect and govern Dataverse solutions—tables, relationships, security roles, business rules, plug‑ins, and elastic tables—serving as the system of record for Copilot Studio agents and Power Apps
• Build agent tools, plug‑ins, and reusable components in C# and .NET 10 to extend Copilot Studio with custom business logic, API integrations, and AI orchestration patterns (tool calling, RAG, declarative agents)
• Operate Power Platform Application Lifecycle Management (ALM) at scale—managed environments, solution packaging, the ALM Accelerator, Power Platform CLI (pac), and pipelines integrated with Git Hub Actions
• Implement responsible AI governance—DLP policies, environment strategy, agent telemetry, evaluation harnesses, content safety, and grounding—ensuring compliance with bank security and regulatory standards
• Partner with software, data, and AI engineers to integrate Azure OpenAI, Azure AI Foundry, and Azure AI Search into Power Platform and Copilot Studio solutions, including knowledge ingestion and retrieval pipelines
• Mentor pro‑code and citizen developers on Copilot Studio patterns, Power Platform ALM, secure coding, and AI engineering best practices; champion compliance with change management and SDLC standards
• It is the responsibility of this role to take ownership of all tasks and challenges that they encounter in the operation of their assigned position. Reasonable accommodations may be made to enable individuals with disabilities to perform the essential functions.

• Bachelor’s degree in Computer Science, Engineering, or related field; or 5+ years of equivalent experience

• 5+ years building enterprise solutions on Microsoft Power Platform (Power Apps, Power Automate, Dataverse), including 2+ years on Microsoft Copilot Studio or comparable conversational AI/agent platforms

• Microsoft Certified:
  Power Platform Developer Associate, Power Platform Solution Architect Expert, Azure AI Engineer Associate, or Microsoft 365 Certified:
  Copilot Studio (one or more preferred)

• Microsoft Copilot Studio: agent design, topics, generative orchestration, knowledge sources, plug‑ins, connectors, authentication, environment promotion, telemetry, content moderation, and evaluation
• Power Platform Connectors: custom connector authoring (OpenAPI definitions), policy templates, OAuth/OIDC, certified vs custom publishing, and connector lifecycle management
• Model Context Protocol (MCP): designing, building, and hosting MCP servers that expose tools, resources, and prompts to Copilot Studio and other AI clients; transport, auth, and tool discovery patterns
• Dataverse design: solution‑aware tables, relationships, role‑based security, business rules, plug‑ins, virtual tables, elastic tables, and integration with external systems
• Agent tooling and patterns: custom plug‑ins, declarative agents, tool calling, retrieval‑augmented generation (RAG), prompt engineering, and agent evaluation harnesses
• Power Platform ALM at scale: managed environments, solutions, ALM Accelerator, Power Platform CLI (pac), pipelines, and tenant‑wide…