Pentesting GenAI LLM models: Securing Large Language Models

Course Overview

• Course Title: Pentesting GenAI LLM models: Securing Large Language Models
• Instructor: Start-Tech Trainings
• Target Audience:
  • Security professionals
  • AI developers
  • Ethical hackers
• Prerequisites:
  • Basic understanding of IT or cybersecurity
  • Curiosity about AI systems and their real-world impact
  • No prior knowledge of penetration testing or LLMs required

Curriculum Highlights

• Key Topics Covered:
  • Unique vulnerabilities of LLMs
  • Penetration testing concepts for generative AI systems
  • Red teaming process for LLMs
  • Traditional benchmarks vs. better evaluation methods
  • Core vulnerabilities (prompt injection, hallucinations, biased responses)
  • MITRE ATT&CK framework for LLMs
  • Model-specific threats (excessive agency, model theft, insecure output handling)
  • Exploitation findings and reporting
• Key Skills Learned:
  • Identifying vulnerabilities in LLMs
  • Red teaming strategies for LLM-based applications
  • Practical experience through real-world attack simulations
  • Applying MITRE ATT&CK concepts for LLM-specific threats
  • Building resilient generative AI applications

Course Format

• Duration: 3.5 hours on-demand video
• Format: Self-paced online course
• Resources:
  • 2 articles
  • 2 downloadable resources
  • Access on mobile and TV
  • Certificate of completion