An AI Engineering Cookbook with 93+ Practical Recipes

What it does

AI Engineering Hub is a curated stack of 93+ Jupyter Notebook projects covering LLMs, RAG pipelines, voice agents, and MCP integrations. The work is sorted into Beginner (22 projects), Intermediate (48), and Advanced (23) tiers, ranging from local OCR apps and ChatGPT clones up to fine-tuning runs and multi-agent deep researchers. Think of it as a self-paced curriculum that happens to live in a GitHub repo.

The interesting bit

Each folder is a vertical slice rather than a loose script: a Hotel Booking Crew or RAG SQL Router is a complete context, not a snippet. The progression from LaTeX OCR with Llama to Build Reasoning Model and an Attention Is All You Need implementation gives you a path from glue code to architecture.

Key highlights

• 93+ projects sorted into Beginner (22), Intermediate (48), and Advanced (23) buckets.
• Heavy focus on RAG variations—basic, agentic, multimodal, voice, and a sub-15ms retrieval stack.
• Explicit MCP (Model Context Protocol) section with Cursor, LlamaIndex, and Firecrawl integrations.
• Model comparison notebooks pitting Llama 4, DeepSeek-R1, Qwen3, Claude, and o3 against each other.
• Includes a full NotebookLM clone and an AI Engineering Roadmap for linear learners.

Caveats

• The README is a high-level index; you have to dive into individual folders to see dependencies or data requirements.
• “Production-Ready” is the repo’s own label—whether each notebook has tests, error handling, or CI is unclear from the top-level docs.

Verdict

Grab this if you learn by breaking working code and need a structured path from simple RAG to agent orchestration. Skip it if you want a single opinionated framework or a managed SaaS; this is a syllabus, not a product.