What you'll learn
Your instructor
Jordan Lee
Agent Systems Engineer, Anthropic
Jordan builds the agent infrastructure behind Claude Code.
Syllabus
When You Need an Agent
Understand the spectrum from single LLM calls to autonomous agents, and know exactly when each pattern is the right choice.
The Agentic Loop
Build your first agent with the Claude Agent SDK and understand the gather-act-verify-iterate loop that powers every autonomous agent.
Designing Tools That Agents Can Actually Use
Design tools that Claude can use effectively — with clear descriptions, mistake-proof schemas, and graceful error handling.
MCP — Connecting Agents to the Real World
Use the Model Context Protocol to give your agent standardized access to external systems — databases, browsers, APIs, and thousands more.
Guardrails, Human Checkpoints, and Multi-Agent Systems
Add safety guardrails with hooks, design human-in-the-loop checkpoints, and orchestrate multiple specialized agents for complex tasks.
Computer Use and What's Next
Give your agent eyes and hands with computer use, see the full agent architecture come together, and learn where the field is heading.
