02 LLM Fundamentals

Phase 1 - Foundations
Stage 2 of 14. Previous: Stage 1 - Prerequisites. Next: Stage 3 - Prompt Engineering.

Goal

Understand the model behavior and API concepts that every agent system depends on.

Learn

• Transformer Models and LLMs
• Tokenization
• Context Windows
• Model Families and Licenses
• Generation Controls
• Pricing and Latency
• Embeddings and Vector Search
• RAG Basics

Build

Create a model comparison notebook or script that tests the same prompt across multiple models and records quality, latency, and cost.

Exit Criteria

• You can explain tokens and context windows.
• You can choose between standard, reasoning, open-weight, and closed-weight models.
• You can estimate cost before running a workload.
• You understand when RAG is better than long context or fine-tuning.

Checkpoint

Use the Stage 2 checkpoint before moving on.