AI Automation Architect (.NET)

We’re looking for an AI Workflow Engineer / Automation Architect to design and build end-to-end AI-enabled business workflows . You’ll productionize solutions using C#/.NET on Azure Functions , integrating LLMs cleanly and safely into existing systems. What You’ll Do Design and implement AI-powered workflows from idea to production Decide what should be handled by LLMs vs deterministic code Build reliable, observable, and scalable automation on Azure Ensure quality, safety, and continuous improvement of AI workflows Must-Have Requirements 1. Strong C# / .NET Engineering Production-grade C# with modern .NET Experience with async programming, dependency injection, error handling, and structured logging Writes maintainable code with unit/integration tests Familiar with CI/CD and code review best practices 2. Azure Functions & Azure Services Hands-on experience with Azure Functions (HTTP, queue, Service Bus triggers) Familiar with Durable Functions (if applicable) Practical experience with: Azure Service Bus Azure Storage Key Vault Application Insights / monitoring Understands configuration, secrets, and scaling concerns 3. LLM & Prompt Engineering Designs prompts with structured outputs (schemas, JSON) Builds guardrails to reduce hallucinations and ambiguity Implements validation, retries, fallbacks, and human-in-the-loop where needed Understands real-world LLM failure modes 4. AI Workflow Orchestration Breaks business processes into clear automation steps Implements orchestration patterns such as: Idempotency Correlation IDs Retries and dead-lettering End-to-end observability 5. Quality & Evaluation Defines evaluation datasets (“golden sets”) Runs regression tests for prompts and workflows Monitors quality drift and improves performance over time Nice-to-Have Power Platform (Power Automate / Power Apps) Document AI / OCR (Azure Document Intelligence or similar) RAG / embeddings and retrieval-based AI solutions Security & governance experience: PII handling Prompt injection awareness Audit trails and data controls