Job highlights

- Recognized cybersecurity certification (e.g.,CISSP,CISM,CCSP,GIAC)
- Basic Qualifications . Proficient in cybersecurity frameworks and standards
- Preferred Qualifications .
- Experience : 5+ years in security architecture

Job_Description: **Duties & Responsibilities

**Define the enterprise AI security architecture & baselines across Azure, GCP, Data-bricks, onu2011prem, third-party vendors, and Microsoft 365 Copilot covering identity, access, data protection, runtime controls, observability, and incident response for internal agentic workflows and customeru2011facing agentic commerce flows.

**Establish agent identity & access guardrails with Microsoft Entra Agent ID/Agent Registry, Conditional Access policies, and leastu2011privilege patterns for both human and machine identities tied to AI agents (including Copilot Studio and the Microsoft 365 Agents SDK).

**Harden model runtime and toolu2011calling using LLM firewalls, prompt injection defenses, content safety, jailbreak protection, and transaction controls aligning GCP Model Armor and Azure Prompt Shields/Content Safety with security telemetry and response playbooks.

**Govern agents of all types , including agent registration, environment segregation, pipelines, sharing policies, DLP integration, and lifecycle controls for maker/owner changes and owner lessagents.

**Secure Data-bricks AI/Mosaic workloads (model endpoints, gateways, Foundry integrations), standardizing authN/authZ, network isolation, secrets handling, and logging across notebooks, jobs, and serving layers.

**Operationalize SOC coverage for AI , integrating signals into SIEM systems,



tuning detections for agent misuse, and using AI agents to accelerate triage, investigations, and incident response.

**Protect AI on endpoints (Windows/macOS) by setting baseline configurations for agentic browsers, extensions, and local AI run times; enforce dueu2011diligence review and hardening before user downloads/enablement.

**Embed privacy, compliance, and disclosure (DLP tools, sensitive data policies) into AI pipelines.

**Architect crossu2011team delivery create reusable patterns, standards, and reference implementations; lead design reviews; coach engineering teams; and drive adoption via roadmaps and enablement.

Requirements

**Basic Qualifications

Proficient in cybersecurity frameworks and standards.

Experienced with cloud security, identity management, threat modeling, and incident response.

**Experience : 5+ years in security architecture; 35+ years designing controls for cloud platforms; handsu2011on with AI/ML security or production LLM workloads.

**Depth in identity & access : Conditional Access, workload identities, service principals, managed identities, and agent identity governance patterns.

**Crossu2011cloud fluency especially with AI PaaS services : Azure, GCP, Data-bricks Mosaic, and Microsoft 365 Copilot/Copilot Studio.





**Threatu2011driven mindset : Familiar with LLMu2011specific risks (prompt injection, data exfiltration, modelu2011assisted fraud, toolu2011calling abuse), secure SDLC for agents, working knowledge of all security control types, and red/blue team techniques adapted to AI.

**Controls & observability : DLP, logging/telemetry, SIEM/SOAR integrations, Zscaler/Defender stack, and incident response playbooks tailored for agentic workflows.

**Leadership & communication : Proven ability to lead architecture programs, influence crossu2011functional stakeholders, write transparent standards, and mentor engineers.

**How youll measure success

**Guardrail coverage : Measurable adoption of baseline controls and policies across all AI platforms and environments.

**Reduced risk : Fewer highu2011severity findings in AI threat modeling, pen tests, and control assessments

**Operational readiness : SOC detections and runbooks for AI threats in place; effective use of SIEM AI tools; auditu2011ready evidence trails.

**Governance maturity : AI Governance tools functioning at scale with clear owner/lifecycle controls.

**Developer enablement : Adoption of secure patterns and reference architectures by product teams; streamlined approvals that balance agility and control.

**Preferred Qualifications

Recognized cybersecurity certification (e.g., CISSP, CISM, CCSP, GIAC).

Cloud security certifications (Azure/GCP)

Familiarity with stateu2011level AI legislation, FTC guidance, and other international AI legislation

,Work_Experience:10-12 years

📌 Principal Architect - AI (Chennai)
🏢 Staples India
📍 Chennai