These are the core competencies that are absolutely essential to deliver the workstreams and deliverables defined.
Area of Expertise
AI & LLM Security The entire project is centered on securing specific Large Language Models (LLMs), making this the most critical domain knowledge required.
Cloud Security Architecture
Workstream 1 is dedicated to designing the core security architecture, which is the foundational blueprint for the entire engagement.
Microsoft Sentinel Expertise
Workstream 3 explicitly requires deep integration with Microsoft Sentinel for all telemetry and threat detection, making this a non-negotiable technical skill.
Data Loss Prevention (DLP)
Workstream 2 focuses on "Guardrail & DLP Engineering," a core implementation phase that requires deep expertise in DLP principles and tools.
SOAR Automation The scope includes delivering "initial SOAR playbooks" as part of the Sentinel integration, requiring the ability to create automated response workflows.
Adversarial AI Validation
Workstream 4, "Adversarial Validation," is a key phase that promises to validate the effectiveness of the controls, requiring offensive security skills focused on AI.
Nice to Have Skills:
These skills would significantly improve the quality, efficiency, and scalability of the project but are not strictly required to meet the core deliverables.
Area of Expertise
Infrastructure as Code (Terraform)
While the scope requires creating "security templates," using IaC like Terraform is a best practice for ensuring consistency and scalability, rather than a hard requirement for the initial build.
Scripting & Automation (Python)
Python is extremely valuable for custom integrations and automation tasks that may fall outside the native capabilities of Sentinel, providing flexibility and power to the team. However, the primary deliverables could be met using the features within the core security tools themselves.