Software engineer → AI engineer
Build LLM apps with Python
An introduction to building applications on large language models, written for software engineers who already know how to code. It takes you from your first API call to an evaluated, cited question-answering system, using Python and Anthropic’s Claude API.
Twelve chapters
- What an AI engineer does
- How LLMs work: tokens, context windows, sampling
- Calling the model API: streaming, stop reasons, retries, cost
- Prompt engineering you can measure
- Structured output and tool use
- Embeddings and vector search
- Retrieval-augmented generation with citations
- Agents and the Model Context Protocol
- Evaluation: knowing it works
- Choosing models and fine-tuning
- Running AI in production
- A capstone project and your career
What you get
Learn by doing
Each chapter has worked Python examples, diagrams and practice steps.
Check what you know
More than 100 quiz questions with explanations, saved best scores, and a shortcut to continue your last unfinished chapter.
Reference when you need it
A glossary, links to the documentation and papers behind each chapter, and a guide to certifications from Anthropic, NVIDIA, Databricks, AWS, Microsoft and Google Cloud.
Run the code
A free companion repository has runnable versions of the named examples, with sample data and offline smoke tests. Run them on a computer or in a browser with GitHub Codespaces.
Private by design
Read, search the glossary and take quizzes offline, with no account, no ads and no tracking. Your progress stays on your device.
Made for iPad too
On larger iPads the chapter list sits beside the chapter you are reading.
Exercises that call the API need internet access, your own Anthropic account and API credits. AI Engineer Path is an independent course. It is not affiliated with or endorsed by Anthropic or any certification provider.