AI - SEC545: GenAI & LLMs App Security (5-Day)

Secure GenAI applications from development through production with comprehensive hands-on training in LLM security fundamentals and advanced defense techniques.

What You Will Learn:

GenAI Security Fundamentals

• Understand LLM architecture, embeddings, and RAG systems
• Identify prompt injection attacks and malicious model risks
• Assess third-party supply chain vulnerabilities
• Secure fine-tuning processes and knowledge sources

Architecture and Deployment Security

• Configure AWS Bedrock security features
• Run local models securely
• Mitigate model hosting supply chain attacks
• Secure agentic AI and MCP server implementations

MLSecOps Integration

• Apply threat modeling with the MAESTRO framework
• Defend against model serialization attacks
• Secure MLOps pipelines and Airflow orchestration
• Implement data protections throughout the lifecycle

The 5-day version expands coverage of agents, MCP, and authentication security. Gain hands-on experience with AI threat modeling, manual model training using Hugging Face Transformers, and a redesigned Capture the Flag focused on GenAI infrastructure.

11 hands-on labs include LLMs and Prompt Injection, Fine-tuning OpenAI Models, Compromising Vector Database, Safe Use and Moderation, AWS Bedrock, Pivoting from LLMs, Compromising LLM Supply Chain, Langchain Security, Model Serialization Attacks, MLSecOps, and a Capture the Flag challenge.

Ideal for Application Security Engineers, Cloud Security Engineers, SOC Analysts, Incident Handlers, Threat Intelligence Professionals, Security Auditors, and Risk Managers.

30 CPE credits. 5 days of expert-led training.