See What You'll Learn in
This Intensive Workshop
Watch our instructor walk through the full AI Security Specialist curriculum, what to expect from each Saturday session, and how the hands-on labs work.
Everything You Need to Master
AI Security — In One Workshop
Offensive AI Labs
Attack AI systems like a red-teamer. Practice prompt injection, adversarial attacks, model poisoning, and LLM exploitation in a real cloud environment.
Defensive AI Labs
Build production-grade defenses: input validation, AI firewalls, threat modeling, incident response, and secure ML pipeline implementation.
Live Instructor Sessions
Every Saturday, 2 hours of live instruction with real-time Q&A. Ask questions, get feedback, and learn directly from an industry expert.
Real Cloud Lab Environment
No setup hassles. Your browser-based cloud lab is ready from day one — attack and defend real AI systems, not simulations.
Session Recordings Included
Miss a Saturday or want to re-watch? Every session is recorded and available to you forever — learn at your own pace.
CAISP™ Certification
Complete all 12 sessions and the capstone project to earn the globally-recognized Certified AI Security Professional™ credential.
Global Cohort Community
Learn alongside security professionals from around the world. Network, collaborate, and grow together in your private cohort community.
Career-Ready Portfolio
Graduate with documented capstone projects that showcase your AI security skills to future employers and clients.
Saturdays — Zero Career Disruption
Learn without quitting your job. Every session is on Saturday for 2 hours — designed around working security professionals.
Extreme Hands-On Experience —
The #1 Reason Students Choose Us
Theory without practice means nothing in cybersecurity. Every module includes real lab exercises where you attack and defend live AI systems in a cloud environment. You leave each session having done the work — not just watched it.
Think like an attacker. Exploit vulnerabilities in AI systems, LLMs, and ML pipelines before your adversaries do.
- Scanning LLMs for agent-based vulnerabilities & attacking AI chatbots
- Performing adversarial attacks using the TextAttack framework
- Prompt injection attacks — bypassing system prompts & security controls
- Training data poisoning & excessive agency exploitation
- Adversarial attacks using Foolbox framework
- Insecure plugin exploitation in LLM ecosystems
- Data leakage exploitation & permission escalation in LLM systems
- Poisoned pipeline attack simulation — compromising ML deployments
- Dependency confusion attacks on package management systems
- Supply chain dependency attack exploitation
- Backdoor attacks using BackdoorBox toolkit
- Red team AI systems using professional offensive methodologies
- Web scraping attacks with PyScrap & steganography using StegnoGAN
- Penetration testing against LLM applications and APIs
Build resilient AI systems. Implement layered defenses that stop real-world attacks at every layer of the AI stack.
- Setup browser-based lab environment & InvokeAI creative visual tool
- Create secure chatbot using Python & machine learning frameworks
- Testing defenses with Adversarial Robustness Toolbox (ART)
- Implementing defensive controls for LLM applications (layered)
- AI threat modeling using STRIDE methodology
- Risk rating workshop — calculating likelihood and impact
- AI threat modeling with IriusRisk & StrideGPT automated tools
- Implementing SCA tools & model scanning for AI security projects
- Generating SBOMs, attestations, and model signing
- Developing comprehensive AI governance framework & compliance
- Adversarial robustness testing & defensive distillation
- Implementing AI-specific Web Application Firewalls (WAFs) & guardrails
- Comprehensive AI security assessment capstone project
- Model watermarking & fingerprinting to detect theft
12 Modules · Scratch to Advanced
A meticulously structured learning path built on real-world AI security scenarios. Each module includes theory, live demonstrations, and hands-on lab exercises.
- AI evolution & security implications across industries
- ML Fundamentals: supervised, unsupervised, reinforcement learning
- Deep learning architectures: neural networks & CNNs
- NLP & computer vision security considerations
- RAG architectures and their security implications
- LLM architectures: GPT & BERT transformers
- MITRE ATT&CK and ATLAS frameworks for AI attacks
- Real-world malicious LLM tools: WormGPT & FraudGPT
- Adversarial attacks using TextAttack framework
- Attacking AI chatbots & scanning LLMs for vulnerabilities
- Complete OWASP Top 10 for LLM applications
- Direct & indirect prompt injection exploitation
- Training data poisoning attacks
- Model denial of service & supply chain vulnerabilities
- Sensitive information disclosure from LLM training data
- DevSecOps principles integrating security into AI development
- Attacks targeting CI/CD pipelines
- Real incidents: Hugging Face breaches & SAP AI Core vulnerabilities
- Software Composition Analysis (SCA) for AI projects
- AI firewalls guarding models against adversarial inputs
- STRIDE methodology for AI system threat modeling
- Creating Data Flow Diagrams (DFDs) for LLM applications
- OWASP LLM, MITRE ATLAS & BIML threat libraries
- AI Risk Repository & AI Incident Database analysis
- Automated threat modeling with IriusRisk & StrideGPT
- AI supply chain attack surface: data, models, infrastructure
- Compromised dependencies & backdoored model analysis
- Vetting processes for third-party AI frameworks
- SLSA & SCVS framework implementation
- Software Bill of Materials (SBOMs) & model signing
- Model-mediated supply chain attacks
- Self-propagating AI model worms
- Backdoors in fine-tuning processes
- AI-assisted evolving firmware & polymorphic malware
- Real-world AI security breach case studies
- NIST AI Risk Management Framework (RMF)
- ISO/IEC 42001 AI management system standards
- EU AI Act requirements & risk classification
- US AI legislation & executive orders
- Comprehensive compliance checklist for AI security
- Input validation & sanitization preventing prompt injection
- Output filtering & content moderation systems
- Access controls & authentication for model APIs
- Model watermarking & fingerprinting
- AI-specific WAFs & guardrails deployment
- FGSM & PGX adversarial example generation
- Adversarial training hardening models against perturbations
- Ensemble methods & model diversity strategies
- Certified defenses providing robustness guarantees
- Continuous monitoring & anomaly detection
- AI-specific incident response procedures & playbooks
- Detecting model poisoning & data exfiltration incidents
- Forensic investigations on compromised ML systems
- Containment strategies for compromised models
- Recovery procedures: model rollback & retraining
- Comprehensive AI security assessment methodology
- Penetration testing against LLM applications & APIs
- Red team operations simulating adversarial attacks
- Automated security testing tools for continuous validation
- CAISP™ certification exam preparation & capstone project
12 Saturdays · 2 Hours Each
Every session runs live from October to December, 2026. Designed for working security professionals — no career disruption required.
All sessions are live on Saturdays — designed for working professionals. Each session runs 2 hours with live Q&A. Recordings are available within 24 hours.