Hands-on, framework-free Colab notebooks teaching the applied-LLM stack for AI Engineers and Forward Deployed Engineers (FDEs). Learn model APIs, RAG, evals, agents, adaptation, and serving using raw API calls instead of wrappers like LangChain or LlamaIndex. Everything runs on the free Groq API (no credit card), with LoRA fine-tuning and self-hosted serving as optional GPU-backed appendices.

Evals are the spine: the "measure before you tune" habit is installed early and returns in every section. Three real case studies include a production support assistant, a pipeline-vs-agent cost showdown, and a red-team robustness benchmark.

Learning order: 00 Setup, 01 Model APIs, 02 Evals I, 03 RAG, 04 Evals II, 05 Agents, 06 Fine-tuning vs LoRA, 07 Serving, 08 Prompt Injection/Security, 09 LLMOps, 10 Customer Craft. Each notebook is self-contained, reads API keys from Colab secrets, and ends with exercises.

Built as the hands-on companion to "Plan: Transitioning to FDE/AI Engineer." OpenAI-compatible throughout, with swappable patterns for Anthropic. Targets backend or full-stack engineers moving into AI roles who can ship production code and want the applied-model layer.