vv176/week1_ai_engineering ? reverse-engineered prompt

Build me a small Python project that demonstrates the basics of working with LLMs and text models in a really practical way.

I want a set of simple scripts and mini demos that show how tokenization works, how to print a vocabulary, how embeddings work, how to call an embeddings API, and how to run a basic GPT 2 style workflow. Please also include a few focused examples that explore sampling, streaming responses, structured outputs, log probabilities, threading, and a simple walkthrough of pretraining, supervised fine tuning, and RLHF at a high level.

Make it easy to run each demo on its own, with clear comments and simple output that helps someone learn by seeing it in action. If anything needs current API details or model behavior, look up the latest docs online first. Keep the code clean, educational, and easy to follow for someone who is just getting into AI engineering.