Applied Agentic AI

3 Credits · Level: Graduate

Applied Agentic AI: Building and Operating Agents - This course teaches AI agents from the ground up. We begin with the foundations an agent builder needs, including how a language model produces tool calls, sampling and context windows, structured output, and retrieval as a tool, then write an agent loop by hand before adopting any framework. We explore key techniques such as agent memory and context engineering, tool use through the Model Context Protocol (MCP), agent-to-agent (A2A) collaboration, agent evaluation, observability, and durable execution for long-running agents. We study the tokenomics of agents, agent safety, agent security and identity, agentic platform design with registries, gateways, and AI traffic management, and autonomous agents that transact under a spend mandate. The course culminates in a capstone project, preparing students to deploy evaluated, observable, and governed agent systems in production.