AI Security Engineer

Pretium Enterprise Services, India•Scottsdale, AZ

1d•Hybrid

About The Position

Pretium Enterprise Services (PES) is seeking a hands-on AI Security Engineer within Infrastructure & Cybersecurity to help build and implement its enterprise AI security program. The role will assess the current security posture, develop a prioritized roadmap, and implement controls that support secure AI adoption across Pretium and its operating companies. The engineer will work with technology teams, developers, business owners, Legal, and Compliance to secure enterprise AI platforms and AI capabilities embedded in business applications. The role will establish shared standards and practical protections for enterprise data, AI agents, and integrations. The position is hybrid, with three days a week in the office. Scottsdale, Arizona is the preferred location; Dallas, Texas, Charlotte, North Carolina, and California are also considered.

Requirements

Five or more years of experience in security engineering, cloud security, application security, or a related discipline, with demonstrated hands-on implementation of enterprise controls.
Practical experience securing an enterprise AI platform or AI-enabled application, including identity, data access, integrations, or monitoring, and testing control effectiveness.
Strong foundations in IAM, least privilege, data protection, API security, secrets management, and security logging, with hands-on AWS or Microsoft cloud experience.
Working knowledge of LLM and agent security risks, including prompt injection, data leakage, excessive permissions, and unsafe tool execution.
Ability to automate controls and integrations using Python, PowerShell, or a comparable language, and work with REST APIs, logs, and connected systems.
Ability to turn evolving requirements into a practical roadmap and deliver work through implementation, documentation, and operational handoff.
Strong communication and collaboration skills across operating companies and control functions.

Nice To Haves

Familiarity with the NIST AI Risk Management Framework and OWASP guidance.
Experience with Claude Enterprise, ChatGPT Enterprise, Microsoft 365 Copilot, AWS Bedrock, Microsoft Entra, Purview, DLP, or SIEM platforms.
Experience with MCP integrations, AI agents, RAG, adversarial testing, or infrastructure as code.
Experience in a shared-services organization, financial services, or another regulated industry.

Responsibilities

Assess AI usage, data flows, and control gaps across operating companies, and maintain an inventory of AI platforms, embedded capabilities, agents, integrations, and accountable owners.
Develop and execute the AI security roadmap, including priorities, milestones, dependencies, and measurable outcomes; establish minimum enterprise controls and repeatable intake, onboarding, review, and exception processes.
Configure and validate platform controls, including SSO, provisioning, role-based access, approved models, sharing restrictions, and connector permissions across web, desktop, developer tools, and APIs.
Implement and tune data loss prevention and AI guardrails; work with data owners to address oversharing, protect sensitive information, and maintain data boundaries between operating companies.
Secure AI agents, plugins, connectors, and Model Context Protocol (MCP) integrations through least privilege, credential protection, restricted tool access, and approvals for high-impact actions.
Partner with cloud and development teams to secure Bedrock and other AI workloads using identity controls, encryption, secrets management, network protections, and automated configuration checks.
Perform threat modeling and security testing for AI applications and retrieval-augmented generation (RAG), including prompt injection, data disclosure, unauthorized actions, retrieval permissions, and guardrail bypass; track remediation and retest changes.
Integrate available audit APIs, cloud logs, and agent telemetry with enterprise monitoring; validate coverage, address gaps, and develop detections and incident response procedures with security operations.
Implement approved retention, access, and deletion requirements with Legal and Compliance, including protection of sensitive information in prompts, outputs, and logs.
Evaluate AI vendors and embedded AI features for data handling, model training use, permissions, retention, and auditability; test security solutions and recommend tools that integrate with enterprise capabilities.
Publish technical standards and implementation guidance, support operating company adoption, and report roadmap progress, control coverage, exceptions, and remediation to leadership.