English
Week 1: Fundamentals
Overview

Week 1: Agent Fundamentals & Tooling

Learn the foundational concepts of AI agents by implementing them from scratch.

Goal: By the end of this week, you'll understand how agents work internally and be able to build basic tool-using agents.

Learning Objectives

  • ✅ Understand the ReAct (Reasoning + Acting) pattern
  • ✅ Implement Thought → Action → Observation loops
  • ✅ Define tool schemas with Pydantic
  • ✅ Use OpenAI Function Calling API
  • ✅ Integrate real web search with Tavily

Sessions

Week Overview

Key Concepts

What is an Agent?

An agent is an LLM that can:

  1. Reason about what to do next
  2. Act by calling external tools
  3. Observe the results and continue

This is fundamentally different from a simple chatbot that only generates text.

The ReAct Pattern

Thought: I need to search for information about X
Action: search["X"]
Observation: [search results]
Thought: Now I have the information, I can answer
Final Answer: Based on my research, X is...

Prerequisites

Before starting this week:

  • Complete the Getting Started guide
  • Ensure your sotaaz environment is active
  • Have your API keys configured

Time Estimate

SessionDuration
Session 1: ReAct~2 hours
Session 2: Tool Calling~2 hours
Weekend Project~3-4 hours

Ready? Let's start with ReAct Pattern →