AI Agents: Foundations & Patterns
Build autonomous AI agents that reason, plan, and act — from the ReAct pattern to tool calling and memory systems. AI agents are systems that perceive inputs, reason about goals, take actions using tools, and iterate until a task is complete.
Level: Intermediate · Category: Agentic AI · Estimated time: 6 hours
Prerequisites
- Machine Learning Basics
- Python for AI
Lessons
- AI Agents & Cognitive Architecture — The perception-reasoning-action loop, how agents differ from chatbots, and the anatomy of an agentic system.
- The ReAct Pattern: Reasoning + Acting — Interleaving thought, action, and observation in a loop — the framework behind nearly every modern agent.
- Tool Use & Function Calling — Defining tools, OpenAI function calling API, schema design, and building a tool-using agent from scratch.
- Memory Systems for Agents — Short-term (context window), long-term (vector store), episodic memory, and conversation summarization.
- Planning & Task Decomposition — Breaking complex goals into subtasks — plan-and-execute, task graphs, and self-reflection loops.
- Building Agents with LangChain — LangChain agents, tools, chains, LCEL, and the AgentExecutor — hands-on from scratch.
- Evaluating & Debugging Agents — Benchmarking agent performance, LangSmith tracing, common failure modes, and adding guardrails.
Topics covered
agents, agentic-ai, langchain, tool-use, react, memory, function-calling, llm-agents