build-llm-applications.md

Build LLM Applications

home

Details

• Title: Build LLM Applications
• Subtitle: (from Scratch)
• Authors: Hamza Farooq
• Publication Date: 2025
• Publisher: Manning
• ISBN-13: 9781633436527
• Pages: 325

Links: Publisher

Blurb

Create your own LLM applications without using a framework like LlamaIndex or LangChain.

In Build LLM Applications (From Scratch), you'll learn to create applications powered by large language models (LLM) from the ground up. In this practical book, you'll build several fully functioning, real-world AI tools—including a search engine, semantic caching for RAG, and autonomous AI agents.

In Build LLM Applications (From Scratch), you'll learn how to:

• Design and implement efficient search algorithms for LLM applications
• Develop custom Retrieval Augmented Generation (RAG) systems
• Master deep customization techniques for every aspect of search and RAG components
• Understand and overcome the limitations of popular LLM frameworks
• Create end-to-end LLM solutions by integrating multiple components cohesively
• Apply advanced fine-tuning techniques for task-specific models and domain adaptation
• Deploy quantized versions of open-source LLMs using vLLMs and Ollama

Build LLM Applications (From Scratch) shows you just how customizable LLM applications can be when you create your own without using opinionated tools like LangChain and LlamaIndex. You'll learn the fundamentals of AI development hands-on, all without any proprietary tools. Soon you'll have the skills you need to build LLM applications, tailor them to your specific needs, and ensure you have control over your entire system. about the book

Build LLM Applications (From Scratch) is a practical and comprehensive handbook for creating custom LLM applications without relying on premade frameworks. You'll start by mastering the fundamentals of search systems and RAG. Then you'll apply this knowledge to real-world projects, including building a hotel search engine using TripAdvisor review data, implementing semantic caching in RAG production systems, and deploying a full RAG application using Hugging Face and Gradio. By the end of the book, you'll have the skills to build AI agents from scratch, deploy open source LLMs with advanced quantization techniques, and create innovative, specialized LLM applications designed for your specific needs.

Contents

PART 1: THE FUNDAMENTALS

• 1. The World of Large Language Models
• 2. An in-depth look into the soul of the Transformer Architecture
• 3. Encoder models in action: Semantic-Based Retrieval Systems

PART 2: RETRIEVAL SYSTEMS

• 4. Semantic Search from Scratch
• 5. Combining Encoder & Decoder Model to Create RAG Applications
• 6. Advanced RAG techniques with knowledge graphs and Semantic Cache

PART 3: BUILDING ENTERPRISE LLM APPLICATION

• 7. Introducing Agents, the next Generation of AI
• 8. Fine-Tuning and Domain Adapation
• 9. Deploying Language Models as APIs
• 10. OpenSource Large Language Models