LLM Security Engineering (Paperback)

Bernard Mallin

ISBN 13: 9798193047025

Published by Independently Published, 2026

About this Item

Paperback. The rush to deploy Generative AI is leaving enterprise applications dangerously exposed. Are your LLMs secure?Traditional AppSec models, WAFs, and standard penetration testing fail when it comes to Large Language Models. In a landscape where natural language is the new attack vector, firewalls cannot stop a cleverly crafted prompt.LLM Security Engineering is the definitive, hands-on guide for security professionals, AI developers, and DevSecOps teams tasked with deploying secure Generative AI in production. Written by an expert AI security engineer, this book bridges the gap between offensive Red Team tactics and defensive engineering, showing you exactly how attackers manipulate LLMs and how to stop them.Instead of theoretical risks, you will dive into practical, repeatable exploit frameworks and robust mitigation strategies. You will learn how to build automated attack surface inventories, harden Retrieval-Augmented Generation (RAG) pipelines, and architect multi-layered output filters that actually work.Inside, you will master: Advanced Prompt Injection: Execute and defend against direct overrides, delimiter breaking, and complex multi-turn conversation poisoning.Indirect Attack Vectors: Secure your system against malicious payloads hidden in retrieved documents (RAG), web content, and tool outputs.Guardrail Bypass & Jailbreaking: Understand token smuggling, persona hijacking, and how to patch vulnerabilities without breaking legitimate user experience.Data Leakage Prevention: Stop training data extraction, system prompt leakage, and PII regurgitation cold.Agent & Tool Security: Sandbox function-calling, scope permissions, and prevent devastating privilege escalation in autonomous AI agents.OWASP GenAI Top 10 Compliance: Map your defenses to industry standards with audit-ready checklists, continuous CI/CD security testing, and automated fuzzing.Whether you are auditing a complex AI agent architecture or building the security review gate for a new enterprise chatbot, this book provides the exact rubrics, test harnesses, and containment procedures you need.Don't wait for a costly data breach or a PR disaster to test your AI's defenses.Scroll up and click "Buy Now" to secure your LLM applications before attackers do. This item is printed on demand. Shipping may be from our UK warehouse or from our Australian or US warehouses, depending on stock availability.

Seller Inventory # 9798193047025

Bibliographic Details

• Title: LLM Security Engineering (Paperback)
• Publisher: Independently Published
• Publication Date: 2026
• Language: English
• ISBN 13: 9798193047025
• Binding: Paperback
• Condition: new

About this title

Synopsis

The rush to deploy Generative AI is leaving enterprise applications dangerously exposed. Are your LLMs secure?
Traditional AppSec models, WAFs, and standard penetration testing fail when it comes to Large Language Models. In a landscape where natural language is the new attack vector, firewalls cannot stop a cleverly crafted prompt.
LLM Security Engineering is the definitive, hands-on guide for security professionals, AI developers, and DevSecOps teams tasked with deploying secure Generative AI in production. Written by an expert AI security engineer, this book bridges the gap between offensive Red Team tactics and defensive engineering, showing you exactly how attackers manipulate LLMs and how to stop them.
Instead of theoretical risks, you will dive into practical, repeatable exploit frameworks and robust mitigation strategies. You will learn how to build automated attack surface inventories, harden Retrieval-Augmented Generation (RAG) pipelines, and architect multi-layered output filters that actually work.
Inside, you will master:
Advanced Prompt Injection: Execute and defend against direct overrides, delimiter breaking, and complex multi-turn conversation poisoning.
Indirect Attack Vectors: Secure your system against malicious payloads hidden in retrieved documents (RAG), web content, and tool outputs.
Guardrail Bypass & Jailbreaking: Understand token smuggling, persona hijacking, and how to patch vulnerabilities without breaking legitimate user experience.
Data Leakage Prevention: Stop training data extraction, system prompt leakage, and PII regurgitation cold.
Agent & Tool Security: Sandbox function-calling, scope permissions, and prevent devastating privilege escalation in autonomous AI agents.
OWASP GenAI Top 10 Compliance: Map your defenses to industry standards with audit-ready checklists, continuous CI/CD security testing, and automated fuzzing.
Whether you are auditing a complex AI agent architecture or building the security review gate for a new enterprise chatbot, this book provides the exact rubrics, test harnesses, and containment procedures you need.
Don't wait for a costly data breach or a PR disaster to test your AI's defenses.
Scroll up and click "Buy Now" to secure your LLM applications before attackers do.