Course Overview: This immersive, 1 day intensive course provides cybersecurity professionals with an in-depth exploration of adversarial tactics, techniques, and jailbreaks (TTJs) related specifically to GenAI. Participants leave able to think and act like a modern GenAI red-teamer, chaining together reconnaissance, prompt manipulation and model-aware exploit generation to achieve real-world impacts, concluding with an advanced, hands-on Vibe Coding Like a Pro session.
Trainer Biography:
Marco Figueroa is the GenAI Bug-Bounty Programs Manager at Mozilla’s 0DIN program, the industry’s first dedicated LLM bug-bounty platform. He leads the global researcher community that dissects guardrails across ChatGPT, Claude, Gemini and open-source LLMs.
Marco’s research has repeatedly shown how hex-encoded and other obfuscated prompts can coerce GPT-4o into writing working exploit code, a technique covered by The Register and Bitdefender’s Hot-for-Security column. He also uncovered the extent of OpenAI’s container file system exposure, demonstrating live upload-and-execute paths inside ChatGPT’s Debian sandbox, as reported in Dark Reading.
Course Modules:
Module 1: Introduction to GenAI Models & their Weaknesses
Module 2: Prompt Engineering Essentials
Module 3: Subverting OpenAI’s Container Environment
Module 4: Direct, indirect & multi-turn injections
Module 5: Bypassing LLM Filters
Module 6: GenAI Generated Exploit Code To 0Day Finder.
Module 7: Vibe Coding Like A Pro
Module 8: Submit Your First GenAI Bug to 0DIN.ai
