AI Powered Course
AI Security Architecture Training
Architect, Secure, and Defend Modern AI-Enabled Environments
12 Dec
Program Highlights
The AI Security Architecture Training is an advanced, hands-on program designed for cybersecurity professionals responsible for architecting, engineering, governing, and securing AI-enabled environments. The course moves beyond introductory AI security concepts to examine how modern AI systems are designed, attacked, tested, and defended. Participants explore real-world attack surfaces across LLMs, agentic systems, RAG architectures, external dependencies, and AI infrastructure while applying practical security engineering, architectural controls, threat analysis, and hands-on testing throughout the program.
16-Hour LIVE Instructor-led Training 
Hands-on Labs Integrated Across Multiple Modules 
Perform AI Threat Modeling and Attack-Path Analysis 
Apply Practical Security Engineering, Architecture, and Defence 
Assess OWASP LLM, Agentic AI, and ML Threats 
Identify Memory, Context, and RAG Abuse Paths 
Analyze Agent Internals and Security Vulnerabilities 
Build Layered Controls for AI Systems 
Test AI Security Controls and Assurance Mechanisms
Training Schedule
- upcoming classes
- corporate training
- 1 on 1 training
| Start - End Date | Training Mode | Batch Type | Start - End Time | Batch Status | |
|---|---|---|---|---|---|
| 12 Dec - 20 Dec | Online | Weekend | 10:00 - 14:00 IST | BATCH OPEN |
Why Choose Our Corporate Training Solution
- Upskill your team on the latest tech
- Highly customized solutions
- Free Training Needs Analysis
- Skill-specific training delivery
- Secure your organizations inside-out
Why Choose 1-on-1 Training
- Get personalized attention
- Customized content
- Learn at your dedicated hour
- Instant clarification of doubt
- Guaranteed to run
About Course
The AI Security Architecture Training is designed for security professionals seeking practical expertise in securing modern AI-enabled environments. The course examines how LLM platforms, agentic systems, RAG architectures, AI infrastructure, and external dependencies introduce new security considerations beyond traditional applications. Participants learn to analyze AI architectures, identify attack paths, assess AI-specific threats, apply security controls, and evaluate system resilience through integrated hands-on labs. The program also covers governance, risk management, compliance, agent vulnerabilities, AI-assisted attacks, testing, assurance, and secure design patterns, helping participants translate cybersecurity principles into practical architecture and engineering decisions for real-world AI systems.
Course Curriculum
-
Module 1: AI Foundations for Architects
- Objectives:Understand enough of how LLMs and AI platforms work to reason about where they break.
- 1.1 Basics of LLMs
- 1.2 Why AI infrastructure is not a traditional application
- 1.3 Anatomy of an AI system
- 1.4 MLOps
-
Module 2: Governance, Risk Management and Compliance (GRC)
- Objectives:Translate AI risk into governance structures, controls, and audit-ready evidence.
- 2.1 Governance structure
- 2.2 Frameworks and regulations
- 2.3 AI risk management
- 2.4 Compliance evidence
-
Module 3: AI Threats
- Objectives:Systematically identify threats to AI systems, including multi-agent designs.
- 3.1 Threat modeling basics and flows
- 3.2 Threat modeling AI systems
- 3.3 Multi-Agent Systems (MAS)
-
Module 4: AI Attack Surface
- Objectives:Know the concrete attack techniques and the architecture-level mitigations.
- 4.1 OWASP Top 10 for LLM Applications, OWASP Agentic AI threats and OWASP ML Top 10 for non-LLM models.
- 4.2 Memory and context abuse
- 4.3 RAG system attacks
- 4.4 MCP and A2A
-
Module 5: AI Defense and Security AI Agents
- Objectives:Build layered defenses and understand where AI can strengthen security operations.
- 5.1 Defense-in-depth for AI systems
- 5.2 Testing and assurance
- 5.3 Security AI agents
-
Module 6: Agent Architecture and Vulnerabilities
- Objectives:Analyze agent internals and design secure agentic systems.
- 6.1 Agent anatomy
- 6.2 Agent vulnerabilities
- 6.3 Secure agent design patterns
-
Module 7: External Dependencies
- Objectives:Manage third-party and supply chain risk across the AI stack.
- 7.1 Dependency map
- 7.2 Risks
- 7.3 Controls
- Note: Labs will be integrated with multiple sections to supplement Theory with practical.
- Lab Prerequisites and Requirements
- System Requirements
- Participants are required to have access to a computer that meets the following minimum specifications:
Component Requirement Operating System macOS, Windows, or Linux CPU 4 or more cores recommended RAM 8 GB minimum; 16 GB recommended Disk Space At least 10 GB of free disk space Internet Access Required for initial software downloads and setup - Note: Admin access to the machine may be needed to run some of the software so personal computers are preferred over corporate machine that may limit access.
Target Audience
This course is ideal for cybersecurity professionals who already understand foundational security concepts and want to develop deeper architectural and engineering expertise for AI-enabled environments.
- Security Architects responsible for designing secure enterprise architectures and integrating AI-related security requirements
- Security Engineers responsible for implementing, testing, and operating security controls and technologies
- GRC Professionals responsible for governance, risk, compliance, security frameworks, and emerging AI-related risks
- Cloud Architects responsible for designing and securing cloud environments supporting AI workloads and services
Pre-requisites
Participants should have a basic understanding of cybersecurity and core security concepts. This is not an introductory cybersecurity or AI course.
Participants should be comfortable with:
- Security architecture and security controls
- Threats, vulnerabilities, and risk
- Authentication and authorization
- Network and application security
- Security monitoring and incident response
The course builds on these concepts to address advanced AI security architecture, attack surfaces, and defensive engineering.
Course Objectives
Upon successful completion of the training, participants will be able to:
- Understand how LLMs, AI platforms, and MLOps environments differ from traditional applications
- Analyze the architecture and security boundaries of modern AI-enabled systems
- Apply governance, risk management, and compliance principles to AI environments
- Conduct threat modeling for AI, LLM, and multi-agent architectures
- Identify AI-specific attack surfaces and emerging adversarial techniques
- Assess risks associated with memory, context, RAG, MCP, and A2A architectures
- Apply defense-in-depth principles across AI systems and infrastructure
- Evaluate security controls through practical testing and assurance activities
- Identify vulnerabilities within agent architectures and workflows
- Apply secure design patterns to agentic AI systems
- Assess external dependencies, third-party components, and AI supply chain risks
- Translate AI security risks into practical architectural controls and defensive strategies
Vision
Goal
Skill-Building
Mentoring
Direction
Support
Success
It was a very good experience with the team. The class was clear and understandable, and it benefited me in learning all the concepts and gaining valuable knowledge.
I loved the overall training! Trainer is very knowledgeable, had clear understanding of all the topics covered. Loved the way he pays attention to details.
I had a great experience with the team. The training advisor was very supportive, and the trainer explained the concepts clearly and effectively. The program was well-structured and has definitely enhanced my skills in AI. Thank you for a wonderful learning experience.
The class was really good. The instructor gave us confidence and delivered the content in an impactful and easy-to-understand manner.
The program helped me understand several areas I was unfamiliar with. The instructor was exceptionally skilled and confident in delivering content.
The program was well-structured and easy to follow. The instructor’s use of real-life AI examples made it easier to connect with and understand the concepts.
Frequently Asked Questions
How does this training help security architects secure AI-enabled environments?
The training helps security architects understand how AI systems differ from traditional applications. It focuses on secure design decisions, trust boundaries, data flows, external dependencies, threat modeling, control placement, defense-in-depth, and architectural patterns for securing modern AI-enabled enterprise systems.
Will this course help me identify AI-specific attack paths?
Yes. Participants learn to identify attack paths across LLM platforms, RAG pipelines, memory and context layers, agents, tools, external dependencies, MCP, A2A, and multi-agent workflows. The course connects these risks to practical architectural controls and defensive strategies.
Does the course cover secure design for agentic and multi-agent systems?
The course covers agent anatomy, agent vulnerabilities, multi-agent system risks, secure agent design patterns, and architecture-level weaknesses that may arise when agents interact with tools, memory, external systems, APIs, users, or other agents.
How does this course address RAG, memory, and context abuse risks?
The course examines how attackers may manipulate retrieval pipelines, memory layers, prompts, context, documents, and connected data sources. Learners understand common abuse paths and how to reduce risk through validation, access control, isolation, monitoring, and secure architecture decisions.
Will I learn how to apply threat modeling to AI systems?
Yes. Participants learn how to apply threat modeling to LLM applications, AI workflows, RAG systems, agentic architectures, and multi-agent environments. The focus is on identifying risks, mapping attack paths, prioritizing controls, and translating findings into actionable architecture improvements.
Does this training help with AI governance, risk, and compliance implementation?
Yes. The GRC module helps learners translate AI security risks into governance structures, risk management practices, control requirements, compliance evidence, and audit-ready documentation. This is useful for organizations adopting AI under regulatory, security, and accountability expectations.
How does this course help secure AI supply chains and external dependencies?
The course covers dependency mapping, third-party components, external services, model dependencies, plugins, APIs, tools, and infrastructure risks across the AI stack. Learners understand how to assess dependency risk and apply controls to strengthen AI supply chain security.
Will I learn how to design defense-in-depth for AI systems?
Yes. The training focuses on layered controls for AI systems, including access control, isolation, monitoring, validation, governance, testing, assurance, secure design patterns, dependency controls, and response readiness. The goal is to reduce risk across the full AI architecture.
How does the course help evaluate AI security controls?
Participants learn how to assess whether AI security controls are practical, testable, and aligned with real attack paths. The course covers testing and assurance approaches that help validate defenses across AI applications, agents, infrastructure, data flows, and external dependencies.
Is this course useful for GRC professionals working on AI risk?
Yes. GRC professionals can use this course to better understand AI-specific risks, control expectations, compliance evidence, AI governance structures, and audit considerations. It helps bridge the gap between AI technical architecture and governance, risk, and compliance requirements.
What practical outcomes can I expect from this training?
Participants will be able to analyze AI architectures, identify AI-specific attack surfaces, conduct threat modeling, assess agentic and RAG risks, apply secure design patterns, evaluate controls, manage dependency risks, and translate AI security concerns into practical architecture and engineering decisions.
