Lesson 1Setting the FoundationFree Preview

Introduction to LLMs & Agents

What are Large Language Models, how do they become agents and how does that change the way we build software?

Overview

This lesson covers the journey from standalone LLMs to autonomous coding agents. We break down how tokens and embeddings work, the tool calling loop, the agentic think-act-observe cycle, MCP and where agents fit across the software development lifecycle.

Visual companion: Follow along with the interactive Miro board as we walk through each concept.

Key Takeaways

  • LLMs predict one token at a time. Temperature controls randomness, not accuracy
  • Tools bridge the gap between reasoning and action. The LLM decides when to call them, your code executes them
  • The agentic loop (think, act, observe, repeat) is the core pattern behind every coding agent
  • MCP standardises how agents connect to external tools and services
  • Agents are powerful collaborators, not replacements. You stay in control
Setting Up Your Environment