A hands-on course covering how large language models work, core prompting techniques, structured outputs and function calling, retrieval-augmented generation, evaluation, guardrails, cost control, and production deployment. Each lesson pairs practical explanations with a concrete action step you can apply to your own projects.
Modules
How LLMs Work
Lessons
Core Prompting Techniques
Lessons
Structured Outputs and Functions
Lessons
Evaluating and Testing Prompts
Lessons
Building LLM Applications
Lessons
Guardrails and Cost Control
Lessons
Shipping to Production
Lessons
Other Lessons
Why Retrieval-Augmented Generation Matters
Chunking and Embedding Your Documents
Retrieval quality is decided mostly before any question is asked, when you split documents into chunks and convert them to embeddings. Poor chunking quietly caps how good your whole RAG system can ever be, so it deserves deliberate choices. Chunking strategies A chunk should contain one self-contained idea, often a few hundred tokens. Splitting on […]
Building and Tuning the RAG Pipeline
With documents chunked and embedded, the online pipeline answers questions in four steps: embed the query, retrieve candidates, assemble the prompt, and generate. Each step has settings worth tuning, and small changes often produce large quality gains. Retrieval settings that matter Decide how many chunks to retrieve, commonly three to ten, balancing coverage against noise […]