Hands-on, framework-free Colab notebooks for learning the applied-LLM stack.
Learn model APIs, RAG, evals, and agents using raw API calls without heavy frameworks.
ai-engineer-notebooks provides runnable Colab notebooks to build working AI systems from scratch. It is designed for backend or full-stack engineers transitioning into AI Engineer or Forward Deployed Engineer roles.
It walks you through prompting, structured outputs, tool calling, RAG, and agent loops using raw API calls. You build patterns from the ground up to understand what wrappers like LangChain actually do under the hood.
It suits backend and full-stack engineers who want to master the applied-model layer of software development. It does not suit non-technical users or product managers looking for a no-code interface.
Write agent loops and RAG pipelines using raw API calls before reaching for heavy wrappers.
Install the habit of measuring outputs with golden sets and LLM-as-judge before tuning.
Run everything end to end on the free Groq API without needing a credit card.
Examine support assistants debugged in production, red-team benchmarks, and compliance systems.
The repository is entirely free under its open-source license, and the notebooks run on the free tier of the Groq API.
879 open remote AI-training roles, updated November 15, 2026; most pay $20–$175 an hour. Zekai lists them free; the application goes to the employer’s platform.
No, everything is designed to run on the free Groq API without requiring a credit card.
Writing systems from raw API calls first helps you understand what wrappers actually do so you can judge when not to use them.
Yes, the patterns are OpenAI-compatible throughout, meaning you can swap the base URL and carry your skills over.
See Zekai first in Google