Express: Microsoft Azure AI Developer

Azure AI developer landscape Developer role and Microsoft Foundry overview GenAI apps, agents, language, and speech App-first vs agent-first solution patterns Boundary with Agent Builder and Cloud Developer Planning and preparing Azure AI development Project setup, resources, environments, secrets Select Azure AI and Foundry capabilities Constraints: data, compliance, latency, cost Build-vs-integrate decision points Selecting, deploying, and evaluating models Foundry model catalog and deployment endpoints Fit by quality, cost, latency, context, and risk Manual and automated evaluation approaches Model decision records and evaluation notes Building generative AI apps Chat application architecture with Foundry Prompts, messages, and system instructions Responses API and maintainable app structure Logging, configuration, and prompt versioning Adding tools and actions Built-in tools, function tools, APIs, custom code Safe tool design, validation, and approvals Tool testing and audit trails When tool use becomes an agent workflow Grounding and RAG Prompt engineering vs RAG vs fine-tuning Retrieval quality, citations, and chunking Foundry IQ and knowledge-enhanced agents Evaluate relevance, groundedness, and safety Developing Azure AI agents Foundry Agent Service and agent architecture Instructions, models, tools, knowledge, state Custom tools and MCP tools where appropriate Agent workflows and business task orchestration Language and speech capabilities Azure Language and text analysis scenarios Speech-capable generative AI applications Translation and voice experiences When specialized services beat a chat pattern Responsible AI, testing, and evaluation Responsible AI in Microsoft Foundry Quality, safety, privacy, and abuse testing Latency, cost, reliability, and UX metrics Deployment-readiness review Developer roadmap to production What belongs in Azure AI Cloud Developer Express Security handoff to Secure AI and Cloud Workloads Data handoff to Fabric Analytics Express Operational checklist for the next sprint