Intermediate → Advanced

IAM, Organizations, logging, detection, data boundaries, Bedrock, and agentic application controls. Built around a production AI workload.

• Design keyless workload identity for AI services
• Collect gate evidence from CloudTrail, Config, Security Hub, and CI
• Apply data-boundary controls for model and retrieval systems

Intermediate

Service account governance, Workload Identity Federation, VPC Service Controls, Vertex AI, audit evidence, and secure data pipelines.

• Remove long-lived keys from AI workload paths
• Build evidence from Cloud Asset Inventory and Audit Logs
• Secure Vertex AI, Gemini, RAG, and data-processing boundaries

All levels

Entra ID, managed identities, Azure Policy, Defender for Cloud, Purview, Key Vault, Private Link, and Azure OpenAI security.

• Design managed-identity access for AI and agent services
• Turn Defender, Resource Graph, and Policy output into audit evidence
• Secure Azure OpenAI and enterprise data access patterns

Intermediate → Advanced

Threat model MCP, tool metadata, memory, identity, action approval, prompt injection, and runtime containment in hands-on labs.

• Build an agent control-plane threat model
• Gate unsafe MCP and tool configurations before deployment
• Instrument traces, approvals, circuit breakers, and evidence records

Leadership + practitioners

A workshop for turning AI security controls into owners, gates, evidence, exceptions, metrics, and a production roadmap.

• Map current AI estate maturity across 8 domains
• Define gate criteria and evidence contracts
• Prioritize runtime, identity, data, and governance backlog items