What you'll learn

AI applications are rapidly evolving from single-shot prompts to complex, autonomous agentic systems. However, most startups and developers struggle with scaling architectures, deploying models efficiently, and monitoring agent behavior in production. In this course, Architecture Patterns for AI Systems, you’ll learn to design and deploy production-ready agentic AI applications using emerging infrastructure and design patterns. First, you’ll explore the layered architecture of agentic systems—covering infrastructure (GPU), model management, vector databases, memory, and popular frameworks like LangChain and LlamaIndex. Next, you’ll discover key design patterns that power today’s most capable agents, including prompt frameworks, RAG, function calling, multi-agent communication, and autonomous decision loops like A2A, ACP, and MCP. Finally, you’ll learn how to serve models efficiently using real-time batch and edge-serving strategies and how to monitor agents in production with AgentOps—tracking key metrics for reliability, performance, and insight. When you’re finished with this course, you’ll have the skills and knowledge of agentic AI system architecture needed to build, deploy, and maintain scalable AI agents for real-world applications.