AI Security Deep Dive (TTAI2800)

AI Security Deep Dive (TTAI2800) is a 3-day, instructor-led training course from IT Dojo covering AI Security Deep Dive (TTAI2800) training and best practices. It is delivered live online or on-site for U.S. federal government, DoD, and corporate teams, employer sponsored.

Course Duration

3 Days

Audience

Employees of federal, state and local governments; and businesses working with the government.

Prerequisites

Ability to read code and understand basic programming concepts, awareness of traditional cybersecurity issues, basic understanding of web applications, familiarity with data handling and basic statistical concepts, and experience with the software development lifecycle and security practices.

Course Description

The AI Security Deep Dive is a comprehensive three-day intensive course designed for programmers, security analysts, and cybersecurity professionals who need to understand and defend against the unique security challenges posed by artificial intelligence and machine learning systems. As organizations increasingly integrate AI into their applications and workflows, the attack surface expands dramatically, creating new vulnerabilities that traditional security approaches cannot address. This expert-led, hands-on course provides the essential knowledge and practical skills needed to identify, assess, and mitigate AI-specific security risks in real-world environments. With a 50 percent hands-on approach, you will master the fundamentals of machine learning from a security perspective, identify and exploit vulnerabilities through guided exercises, and implement practical defenses against data poisoning, adversarial attacks, and privacy breaches. Days 1 and 2 feature interactive labs delivered via Jupyter notebooks, and Day 3 focuses on real-world integration, exposing local models via a Flask API and integrating with a Large Language Model using the Hugging Face Inference API. Integration labs offer multiple language options including Python/Flask, Java/Spring, ASP.NET, and Node.js.

Learning Objectives

• Master AI/ML security fundamentals, attack vectors, and the AI attack surface.
• Identify and exploit AI-specific vulnerabilities through hands-on exercises.
• Implement comprehensive defenses against AI security threats.
• Secure traditional applications that integrate AI models and LLM APIs.
• Protect sensitive information in AI training and inference.
• Establish enterprise-grade AI security governance and incident response.

Course Outline

Day 1: AI/ML Foundations and Attack Fundamentals

• AI/ML security foundations and the OWASP Top 10 as it applies to AI-powered applications.
• Essential AI/ML concepts for security professionals: supervised vs unsupervised learning, neural networks, deep learning.
• AI system architecture, deployment patterns, and the AI threat landscape.
• Data poisoning and training attacks: targeted, untargeted, clean-label, and backdoor attacks.
• Lab: Setting up an AI security testing environment; implementing and detecting data poisoning.

Day 2: Adversarial Attacks and Model Security

• Adversarial examples: perturbation-based and gradient-based methods (FGSM, PGD).
• Model evasion: black-box vs white-box attacks; transferability and physical-world attacks.
• Privacy attacks: membership inference, model inversion, property inference, and model stealing.
• Differential privacy fundamentals and implementation strategies.
• Lab: Generating adversarial examples; conducting membership inference and model inversion.

Day 3: Secure AI Integration and Enterprise Defense

• Secure API integration patterns: authentication, rate limiting, and input validation.
• LLM integration security: prompt injection defenses, output validation, and context isolation.
• Enterprise AI security strategy: governance, monitoring, red teaming, and incident response.
• Advanced and emerging threats: LLM jailbreaking, multi-modal attacks, and supply chain security.
• Lab: Secure LLM integration via the Hugging Face Inference API; building input validation pipelines.

Frequently Asked Questions

Official Certification & Standards Resources

Authoritative references for AI Security Deep Dive (TTAI2800) from the certifying bodies and federal standards that govern this training:

• NICCS Training Catalog (CISA)
• NICE Cybersecurity Workforce Framework (NIST)