Your cart is empty.
Keep shopping
Full MasterClass AI, RAG, Jaibreak Red Teaming 2026
207 students
What you'll learn
- Execute advanced prompt injections and persona jailbreaks to reliably bypass LLM safety guardrails.
- Exploit RAG pipelines and autonomous AI agents using zero-click document poisoning and SSRF attacks.
- Apply adversarial machine learning to perform FGSM, PGD, and black-box evasion attacks on AI models.
- Conduct end-to-end AI red team operations using the C2C framework and implement robust AI defenses.
Explore related topics
Course content
6 sections • 16 lectures • 3h 20m total length
Requirements
- Basic Python programming experience. You should know how to write simple scripts, use variables, loops, functions, and install packages using 'pip'. No ML PhD required!
- Basic comfort with the Command Line / Terminal (CMD, PowerShell, Bash) to run Python scripts, navigate folders, and install open-source tools.
- A fundamental awareness of AI concepts (knowing what "training" and "inference" mean). We will teach you the advanced AI architectures from the ground up.
- A standard computer (Windows, Mac, or Linux). No expensive GPU is required! All 28 hands-on labs can be run entirely for free in the cloud using Google Colab.
Description
Course Description:
Artificial Intelligence and Large Language Models (LLMs) are being deployed at breakneck speed across every industry, but traditional cybersecurity methodologies are failing to secure them. You cannot use standard penetration testing tools to find behavioral vulnerabilities, hallucination exploits, or neural backdoors.
Welcome to the Full MasterClass AI, RAG, Jailbreak Red Teaming 2026—the most comprehensive, hands-on guide to offensive AI security available today.
In this cutting-edge masterclass, you will step into the shoes of a professional AI Red Teamer. Moving far beyond basic "prompt tricks," this course dives deep into the technical execution of modern adversarial AI attacks. Across 11 intensive modules and 28 fully practical, code-driven labs, you will learn exactly how to break, manipulate, and ultimately secure production-grade AI systems.
What You Will Master:
Prompt Injection & Jailbreaking: Bypass strict RLHF safety alignments using advanced techniques like Many-Shot targeting, fiction framing, token smuggling, and multi-modal (vision) injections.
RAG & Vector DB Exploitation: Exploit Retrieval-Augmented Generation pipelines. Learn to execute zero-click document poisoning, manipulate vector embeddings, and trigger Agentic SSRF (Confused Deputy) attacks to exfiltrate private data.
Adversarial Machine Learning: Dive into gradient math to execute FGSM, PGD, and Carlini & Wagner evasion attacks. Force image and text classifiers to confidently make the wrong predictions using IBM’s Adversarial Robustness Toolbox.
Data Poisoning & Backdoors: Corrupt the AI factory. Plant hidden "sleeper agents" in fine-tuning datasets and flip labels to drastically degrade model accuracy.
Model Theft & Supply Chain Attacks: Steal proprietary model weights via API extraction, recover sensitive training data through membership inference, and execute Remote Code Execution (RCE) using malicious PyTorch Pickle files.
The C2C Methodology:
This course introduces the exclusive C2C (Concept → Chain → Compromise) framework. You won't just learn isolated tricks; you will learn how professional red teams chain multiple minor vulnerabilities into catastrophic zero-click exploits. You will build a state-of-the-art local lab utilizing industry-standard tools including NVIDIA Garak, Microsoft PyRIT, Promptfoo, Ollama, and TruffleHog.
By the end of this course, you will know how to scope a professional AI engagement, calculate custom CVSS-AI severity scores, deliver actionable reports, and implement the robust "4-Gate Defense Architecture" to protect against the exact attacks you just performed.
Whether you are a penetration tester looking to future-proof your career, a machine learning engineer securing your proprietary models, or a developer building LLM-driven applications, this masterclass will give you the highly sought-after skills needed to thrive in the AI era.
Enroll today to start breaking the machine so you can learn how to secure it!
Who this course is for:
- Penetration Testers, Ethical Hackers, and Red Teamers looking to future-proof their careers by mastering offensive AI security and LLM vulnerability assessments.
- AI/ML Engineers and Software Developers building LLM, RAG, or Agentic AI applications who need to understand attacker methodologies to build secure, hardened systems.
- AppSec and DevSecOps Engineers responsible for auditing generative AI applications, ensuring OWASP LLM Top 10 compliance, and integrating automated AI scanning into CI/CD.
- Security Architects and Tech Leads who need a deep technical understanding of AI attack surfaces, MITRE ATLAS, and the NIST AI RMF to safely deploy enterprise AI.