Course Objective & Scope
Isolated jailbreak attempts are not enough to show how a production AI application will behave against a determined adversary. This course presents a full AI red teaming methodology spanning scope and rules of engagement, target reconnaissance, multi-turn conversation manipulation, indirect injection through RAG and exploitation of agent tool chains.
Participants combine automated attack orchestration using Microsoft PyRIT, NVIDIA garak and promptfoo with manual creative exploitation, leaving with a reusable red-team test set and finding report template tailored to their organisation.
Who Should Attend?
- Penetration testers and enterprise red team operators
- Application security (AppSec) and AI security engineers
- AI platform teams validating LLM and agent systems prior to production
- Prerequisite: Python scripting and foundational web/API penetration testing skills
Key Takeaways
- Define scope, success criteria and rules of engagement for an AI red team exercise
- Perform reconnaissance across a target's model, system prompt, RAG and tool surfaces
- Orchestrate automated multi-turn attack campaigns with PyRIT and garak
- Develop hybrid (manual + automated) exploit chains that bypass guardrail layers
- Prioritise findings in CVSS/OWASP context and convert them into remediation regression tests
Curriculum
AI Red Teaming Methodology and Planning
- Contrasting classical red teaming, AppSec pentesting and AI safety vs security red teaming
- Defining scope, threat actor profiles and rules of engagement
Reconnaissance and Attack Surface Mapping
- Model family fingerprinting, system prompt probing and guardrail behaviour analysis
- Enumerating connected data sources, tool schemas and error disclosures
Automated Attack Orchestration (PyRIT and garak)
- Scaled probing with garak and PyRIT's converter/orchestrator architecture
- Using an attacker LLM to drive multi-turn (crescendo, tree-of-attacks) campaigns
RAG and Agent Exploitation Chains
- End-to-end exploitation combining poisoned content, unauthorised retrieval and tool chaining
- Lab on guardrail evasion via encoding, language switching and payload splitting
Reporting, Reproducibility and the Purple Team Loop
- Documenting probabilistic findings with Attack Success Rate (ASR) evidence
- Turning red team findings into CI/CD regression suites with blue and engineering teams
Other Courses in This Category
-
Teaches how to design and operate low-latency guardrails across input, retrieval, tool calling and output stages using NeMo Guardrails, Guardrails AI, Llama Guard and deterministic policy engines.
- Live Course
- 5 modules
-
Teaches how to build automated security evaluation suites for LLM and agent applications using promptfoo, garak and Inspect AI, enforce regression gates in CI/CD pipelines and continuously test model and prompt changes.
- Live Course
- 5 modules
-
Teaches how to design enterprise LLM, RAG and agent platforms on Zero Trust principles using a central AI gateway, network and tenant isolation, KMS key management, secure MCP integration and auditable telemetry.
- Live Course
- 5 modules
-
Comprehensive training on PII masking, reversible pseudonymisation, anonymisation and data residency architectures to adopt LLMs without exposing personal or confidential corporate data.
- Live Course
- 5 modules
-
Hands-on training on building a low-latency PII detection, masking and de-pseudonymisation gateway in front of LLM APIs using Microsoft Presidio, GLiNER and Turkish rule/NER recognisers.
- Live Course
- 5 modules
-
Teaches how to replace production customer data in testing, analytics and LLM evaluation environments using k-anonymity, differential privacy and SDV/LLM-based synthetic data generation.
- Live Course
- 5 modules
Get in touch
