From Design to Deployment - Securing AI Architectures

With AI agents and LLMs moving into production, a new attack surface has emerged that demands more than just prompt filtering. True AI security requires a deep understanding of the entire ecosystem, including foundational models, the supply chain for third-party components, vector databases, and the agentic frameworks that grant them functionality.

Hear from Immersive Principal Security Engineer, Ashley Kingscote and Principal Application Security SME, Chris Wood, for a technical breakdown of modern AI architecture and a practical, engineering-first guide to mitigating these complex threats.

You'll hear about:

• Deconstructing the AI System: Move beyond high-level concepts and see a practical breakdown of AI application architecture, including the roles of Large Language Models (LLMs), Retrieval Augmented Generation (RAG), and Model Context Protocol (MCP).
  
  
• The OWASP Top 10 for LLMs: Dive deep into the OWASP Top 10 for LLMs, exploring critical risks like Prompt Injection, Data Poisoning, Excessive Agency, and Supply Chain Vulnerabilities with developer-focused analogies and examples.
  
  
• The Lethal Trifecta: Understand the dangerous intersection of AI systems having access to private data, the ability to communicate externally, and exposure to untrusted content—and why securing this trifecta is paramount.
  
  
• A Lifecycle Approach to Defense: Learn how to embed security throughout the entire AI development lifecycle, from secure design and threat modeling to robust deployment and operational monitoring in cloud environments.