Security for AI Systems: Certified Practitioner Program

This is a synchronous e-Learning program aimed at equipping individuals with practical Blue-team and Red-team skills for safeguarding AI/LLM systems. The program covers both theory and hands-on practice sessions across the following topics:

1. Introduction: How Security for AI systems differs expands on traditional application and network security

2. Deep dive on OWASP LLM Top-10 and beyond

3. AI model basics

• Opensource models and model deployment
• Vector databases and embeddings
• Langflow and Langchain for chat agent development

4. Red Team perspectives:

• Pyrit, Kali Linux and hacker tools
• OSINT for AI systems
• Model theft and data extraction
• Training data poisoning
• Prompt injection techniques

5. Blue team perspectives

• Model vulnerability scanning
• Guardrail models
• Watermarking
• Application security