Building effective LLM-based applications with Semantic Kernel (en Inglés) - Willem Meints

Drawing from the author's trial-and-error experiences, he walks you through proven patterns and workflows that work, from basic prompt engineering to building agents.

You'll learn how to enhance LLMs with external tools, implement RAG for grounding responses in your data, and orchestrate sophisticated AI workflows. Through practical examples and case studies, you'll master the essential techniques for building reliable, scalable AI applications that solve real business problems.

Whether you're adding AI capabilities to existing enterprise software or building new AI-native applications, this book provides the concrete patterns and battle-tested approaches you need to succeed with LLMs.

Working with different LLM providers and designing around brittle API endpoints are challenging problems on their own. Semantic Kernel provides a way to abstract away from the models and API endpoints and start thinking about smart application building blocks. In this book, you learn useful patterns and practices to get the most out of Large Language Models without solving all the challenging low-level problems.

This book is for C# developers and software architects who want to use an LLM in their application to solve specific challenges that can't be solved with normal program logic.

Understanding Large Language Models

Essential LLMOps knowledge

Getting Started with Semantic Kernel

The art and nonsense of prompt engineering

Testing and monitoring prompts

Enhancing LLMs with tools

Retrieval augmented generation

Working with structured output

Prompt chaining workflows

Intelligent Request Routing Workflows

Working with agents