AI Security Architect role

Job Description

Microsoft Purview DSPM — the rarest and most decisive filter; hands-on DSPM for AI experience separates real candidates from label-only resumes.

AI Security Architecture (agentic AI / LLM security) — the core mandate; catches people securing Copilot, agents, and AI workloads.

Azure Security (Sentinel, Entra, Defender for Cloud) — the required stack baseline.

SC-100 / CISSP — the certification shorthand for a senior Microsoft cybersecurity architect.

Multi-cloud security (Azure + AWS) — filters for those who can also secure Bedrock/Lambda/IAM, per the

For this position, the customer is seeking candidates with hands-on AI security architecture experience, specifically professionals who have designed, implemented, or secured AI/ML platforms, rather than candidates whose experience is primarily focused on data security posture management. Based on our review, the profiles shared do not clearly demonstrate or explicitly highlight the required AI security architecture experience.

Please prioritize candidates with expertise in the following areas:

LLM application security

ML pipeline and MLOps security

AI gateways, guardrails, and governance controls

Threat modeling aligned with the OWASP LLM Top 10 and MITRE ATLAS frameworks

Enterprise AI reference architecture and security control design

Own the end-to-end security architecture for Client's Agentic AI platform and carry the data-security mandate for AI. This senior role consolidates two Critical scopes — agentic AI security architecture and Data Security Posture Management (DSPM) — into a single accountable owner who designs the secure platform and protects the data flowing through it.

Key Responsibilities

•        Define and own the agentic AI reference security architecture (v1 and beyond), including Platform 1.0 hardening.

•        Stand up and operate Microsoft Purview DSPM for AI, accelerated via Microsoft FastTrack.

•        Establish the sensitivity-label and ML auto-classification baseline; map and govern data flows into and out of agents.

•        Set architecture standards for agent-to-API and agent-to-data access patterns.

•        Design and secure agentic AI workloads across a multi-cloud estate spanning Azure and AWS, ensuring consistent controls regardless of where 

         agents are built and run.

•        Partner with the identity, AppSec, and GRC functions to ensure controls are designed in, not bolted on.

Required Skills & Qualifications

•        Deep Azure / M365 security architecture experience; Microsoft Sentinel, Entra, Defender for Cloud.

•        Hands-on understanding of how to build and secure agents in a multi-cloud setup, including AWS (e.g., Bedrock Agents, AWS Lambda, IAM) alongside Azure.

•        Hands-on Microsoft Purview — DSPM for AI, sensitivity labels, data-flow mapping, ML-based auto-classification.

•        API and agent architecture security; familiarity with Copilot Studio / agent platforms.

•        Proven track record designing enterprise-scale secure architectures in a regulated environment.

Preferred Qualifications

•        Experience in life sciences, medical device, or other regulated manufacturing.

•        Relevant certifications (e.g., CISSP, Azure Security Engineer, SC-100).

What This Role Unlocks

A secure agentic platform with both Critical gaps closed — agent architecture and data security for AI.