Overview
Getting an LLM to answer a question is a basic beginning, not a finished application. The useful parts come next: connecting it to your data, giving it the ability to use tools, and making the entire system do something actually worth doing.
By the best selling author of Ethical AI and Agentic AI Architecture, The AI Cookbook is a hands-on guide to building software with LLMs and open-source agents. Written for working programmers, it takes you from your first model call to complete applications you can run, inspect, test, and adapt. You don't need to train your own model or become a machine-learning researcher to get AI to solve real user requirements.
The recipes tackle practical jobs: extracting structured information, writing and summarizing, searching private knowledge, querying data in natural language, and working with documents, images, and voice. You'll also build assistants that take action, agents that research and monitor, and applications that incorporate an existing open-source coding agent.
This is not just about asking AI to write your software. It is about making AI a useful, controlled part of the software you deliver.
Along the way, you will learn to:
- connect model output to application logic through validation and explicit contracts;
- retrieve relevant information and check whether answers are supported by it;
- give agents bounded permissions, recover from failures, and keep people in control;
- and evaluate quality, cost, and behavior before and after deployment.
The book provides complete runnable projects, not isolated snippets with the difficult parts left as an exercise. Setup, tests, failure cases, and adaptation guidance are part of the recipe. You'll see what happens when an answer is wrong, a tool fails, or a workflow needs human review-not just what happens when everything cooperates.
A practical companion to Agentic AI Architecture and Ethical AI, this book turns principles into implementations. It is for developers, technical leads, and teams who want to add useful AI capabilities to their own applications without mistaking a convincing demonstration for a dependable product.
An impressive LLM response is an ingredient, not a finished application. This book shows you how to turn it into software worth using.
Customers Also Bought
Details
- ISBN-13: 9798182898669
- ISBN-10: 9798182898669
- Publisher: Agilistic AI
- Publish Date: October 2026
- Dimensions: 9 x 6 x 0.74 inches
- Shipping Weight: 1.05 pounds
- Page Count: 356
Related Categories
