This directory contains examples demonstrating various principles and best practices in working with Generative AI, LLMs and Agentic AI.

This example shows how to enforce structure and reliability in LLM outputs by: Using Pydantic models to define expected output schemas Validating LLM responses to ensure type safety and consistency Handling unstructured natural language responses by converting them into typed Python objects Providing clear failure modes when outputs don't match expectations This is useful for scenarios where downstream systems depend on predictable, structured data from LLMs.