AI Agents
Design agents that reason, use tools and act: agent loops, tool use, memory, planning, multi-agent systems and agent frameworks.
What Are AI Agents?
Autonomous systems that perceive, reason, act and observe — the agent loop explained.
Agent Architecture
The anatomy of an agent: model, tools, memory, planner and executor.
Tool Use & Function Calling
Give agents real capabilities with typed tools and safe execution.
Reasoning & Planning (ReAct)
Interleave thought, action and observation — the ReAct pattern that powers modern agents.
Agent Memory Systems
Working memory, long-term memory and retrieval — state that persists across turns.
Multi-Agent Systems
Specialized agents that collaborate: orchestrators, workers and handoffs.
Building Agents with LangChain
Compose agents with the LangChain ecosystem: tools, memory and executors.
Building a Research Agent
A practical build: gather sources, synthesize findings, cite everything.
Browser Automation Agents
Agents that browse the web: navigation, form filling and screenshots.
Coding Agents
Agents that read, edit and run code: repo context, patch generation and verification.
RAG Agents
Agents that decide when and what to retrieve — retrieval as a tool, not a fixed step.
Evaluating Agents
Measure agent quality: task success, trajectory quality and cost.
Agent Safety & Control
Constrain autonomy: limits, approvals, sandboxes and fail-safes.
Human-in-the-Loop Patterns
Design agents that escalate and collaborate with people instead of acting alone.
Agent Observability
Trace every thought, tool call and step — because agents fail in the middle, not the end.
Deploying Agents
Run agents as services: queues, timeouts, scaling and versioning.
Agent Design Patterns
Reusable blueprints: reflection, tool use, planning, multi-agent collaboration.
Enterprise Agent Applications
Support, HR, finance and knowledge agents inside real organizations.
Agent Cost & Scale
Make agents affordable: token budgets, model routing and parallel workers.
The Future of Agents
Long-horizon tasks, self-improvement and the research frontier.
AI Agents Roadmap
Synthesize the course into a build plan: ship agents, measure them, iterate.