Welcome to this interactive, hands-on guide to building AI-powered applications for business. Every chapter is a runnable Jupyter notebook — you learn by doing, not just reading.
What this book covers¶
This book takes you from zero to deploying AI applications, organized into six parts:
Part I — Foundations: How LLMs work under the hood (tokenization, embeddings, attention) and how to call them via APIs with automatic provider fallback.
Part II — Retrieval-Augmented Generation (RAG): The core technique for grounding LLM answers in your own data. We build RAG pipelines from scratch, then explore variants: vector databases (ChromaDB), SQL databases, graph databases, and even image/video retrieval.
Part III — RAG Frameworks: The same RAG concepts implemented with production frameworks — LlamaIndex, LangChain, and LangGraph — so you see both the manual and the automated approach.
Part IV — Fine-Tuning: When RAG isn’t enough, QLoRA lets you bake knowledge directly into model weights on a free Colab GPU.
Part V — Agents: LLMs that decide which tools to call — from single-agent systems to multi-agent orchestration to autonomous planning workflows. Includes a real-world example (KU Parking Assistant) with tool calling via MCP.
Part VI — Deployment: Taking your AI from a notebook to production — MCP servers on Google Cloud Run, no-code workflows with Dify, and the llm_cascade package for provider management.
How to use this book¶
Every chapter is a Jupyter notebook. You can:
Read online — browse the rendered HTML right here
Run in Colab — click the 🚀 rocket icon at the top of any page to open it in Google Colab
Run locally — clone the source repo and run notebooks in VS Code or JupyterLab
All notebooks use the llm_cascade package for automatic LLM provider fallback. Set any one of 8 API keys (Gemini, OpenAI, Groq, etc.) and every notebook works out of the box.
How to cite this book¶
Srinivasan, K. (2026). AI for business: A hands-on guide. https://
karanalytics .github .io /AIHandsOnBook/
BibTeX:
@book{srinivasan2026aiforbusiness,
title = {AI for Business: A Hands-On Guide},
author = {Srinivasan, Karthik},
year = {2026},
url = {https://karanalytics.github.io/AIHandsOnBook/}
}Author¶
Karthik Srinivasan, University of Kansas School of Business.
A Note on AI Collaboration¶
This book was developed through an extensive human-AI collaboration. The author used Claude Code (Anthropic) as a development partner throughout the process — from generating and debugging the Python notebooks, to building the llm_cascade package, to drafting chapter prose, to assembling the book infrastructure. The AI assisted with code generation, content synthesis from lecture materials, technical writing, and iterative debugging. All content was reviewed, directed, and approved by the author, who takes full responsibility for the accuracy and pedagogical choices in the final work.
In many ways, the creation of this book is itself a demonstration of the AI-human collaboration patterns it teaches — particularly the vibe coding workflow described in Chapter 7 and the agent-assisted development covered in the later chapters.
Built with Jupyter Book. Source notebooks at github