Modern AI Engineering
Build real, production-minded features on top of large language models: prompting, structured output, embeddings, retrieval, tool calling, evaluation, and shipping responsibly.
0 of 10 lessons complete
About this course
A practical, 2026-relevant course for developers who want to build AI-powered features. You will learn what LLMs actually do under the hood, how tokens and context windows work, how to prompt reliably, how to get structured JSON output, how embeddings and similarity search work, how retrieval-augmented generation (RAG) is built, how tool/function calling lets models act on your data, how to evaluate AI output, how to manage cost/latency/limits, and how to ship AI features responsibly.
Syllabus
Foundations of Language Models
What LLMs are, how they represent text, and how to prompt them well.
Foundations of LLMs
How large language models actually work under the hood.
Structuring & Retrieving Knowledge
Getting reliable structured output and grounding models in real data with embeddings, RAG, and tools.
Working With AI Output
Prompting, structured output, embeddings, and retrieval.
Prompting & Structured Output
Writing effective prompts and getting reliable structured results.
- Prompting FundamentalsUnlimitedFinish the lesson, Pass the quick check and pass the coding challenge in "Tokens and Context Windows" to unlock this lesson.12m
- Structured Output and JSON SchemasUnlimitedFinish the lesson, Pass the quick check and pass the coding challenge in "Prompting Fundamentals" to unlock this lesson.12m
Building Reliable AI Products
Evaluating AI output and shipping AI features that are fast, affordable, and safe.
Building Real AI Features
Tool calling, evaluation, cost, and responsible shipping.
- Evaluating AI OutputUnlimitedFinish the lesson, Pass the quick check and pass the coding challenge in "Prompting Fundamentals" to unlock this lesson.14m
- Cost, Latency, and LimitsUnlimitedFinish the lesson, Pass the quick check and pass the coding challenge in "Tokens and Context Windows" to unlock this lesson.13m
- Shipping an AI Feature ResponsiblyUnlimitedFinish the lesson, Pass the quick check and pass the coding challenge in "Evaluating AI Output" to unlock this lesson.14m
- Build a Grounded AI Assistant PipelineUnlimitedComplete "Shipping an AI Feature Responsibly" to unlock this project.~6h
Retrieval & Tools
Grounding models in real data and letting them call functions.
- Embeddings and SimilarityUnlimitedFinish the lesson, Pass the quick check and pass the coding challenge in "Structured Output and JSON Schemas" to unlock this lesson.14m
- Retrieval-Augmented GenerationUnlimitedFinish the lesson, Pass the quick check and pass the coding challenge in "Embeddings and Similarity" to unlock this lesson.14m
- Tool and Function CallingUnlimitedFinish the lesson, Pass the quick check and pass the coding challenge in "Retrieval-Augmented Generation" to unlock this lesson.14m
Shipping AI Responsibly
Evaluating quality, managing cost and latency, and shipping safely.