Principal IAM AI Engineer

About the role

Amex GBT is a place where colleagues find inspiration in travel as a force for good and – through their work – can make an impact on our industry. We’re here to help our colleagues achieve success and offer an inclusive and collaborative culture where your voice is valued. We are seeking a Principal IAM AI Engineer who brings an AI-first mindset rather than a traditional security-first lens to identity architecture. This role sits at the intersection of IAM, AI/agentic systems, and product thinking, for someone who's as comfortable designing a scalable access model for AI agents as they are challenging why a control exists in the first place. It also sits at the intersection of engineering, risk management, and the technical realities of running global travel technology: booking platforms, payment flows, traveler data protection, and extensive supplier integrations. We need someone who can look at how AI agents, copilots, and autonomous workflows are being adopted across the business, understand the actual product and business intent behind them, and build security and identity architecture that enables adoption safely rather than just implementing it. This is a technical and hands-on engineering position where you will construct the authentication and authorization infrastructure, credential management systems, and compliance frameworks that enable responsible AI deployment across the organization. Additionally, you will lead a technical team to maintain, expand, and evolve these capabilities. You will be integrated into AI initiative planning from inception, ensuring identity, access controls, and secrets management are embedded in the architecture from day one rather than implemented after deployment.

Location: United States

Responsibilities

• Access control execution. Deploy and automate identity controls at runtime for NHI identity — including immediate authentication verification, access authorization, and real-time rule application across the environment.
• Machine identity & agentic system governance. Build, deploy, and automate lifecycle governance for agents and machine identities — spanning discovery and registration, permission assignment, credential lifecycle, periodic access reviews, and secure retirement.
• Credential workflows for AI workloads. Construct and operationalize credential management systems for AI-driven applications and agents, encompassing automated secret cycling, short-lived credential issuance, and protected distribution to runtime environments.
• Granular authorization & declarative policies. Establish and implement fine-grained, zero-standing-privilege access models for agents and workloads, documented and managed through infrastructure-as-code and policy declaration frameworks.
• Integration with AI development lifecycle. Collaborate with application development and ML teams throughout all AI initiative phases, ensuring identity requirements, credential handling, and access controls are incorporated during design rather than retrofitted.
• End-user identity controls rollout. Facilitate implementation and adoption of identity governance controls for human users as required by organizational standards.
• Compliance, audit & risk alignment. Work with compliance, risk and audit functions to confirm controls align with regulatory requirements and internal policies; facilitate audit evidence generation and regulatory reporting.
• Organization building & technical strategy. Lead an engineering team supporting these initiatives; establish technical direction, strategic priorities and delivery roadmap for machine identity and agent IAM programs.
• Advisory, enablement and leadership. Serve as a trusted advisor to senior leadership on machine and AI identity risk, and translate it into funded, sequenced remediation. Partner with InfoSec and AI Security teams on proof-of-concept evaluations and vendor assessments for AI and agent governance platforms. Mentor engineers and architects on non-human and agent identity patterns, and raise the organization’s overall fluency in machine identity security.

Requirements

• Deep expertise in enterprise identity infrastructure: PAM (CyberArk or equivalent), Okta as IdP, and IGA platforms like Saviynt, including extending governance beyond human identities.
• Strong AWS and cloud identity skills: IAM design, ephemeral credentials, OIDC federation, secrets/certificate management, and least-privilege at scale.
• Mastery of core identity standards (OAuth, OIDC, SAML, SCIM, JWT, mTLS) and machine-to-machine/workload identity across on-prem, hybrid, and multi-cloud environments.
• Proven ability to architect enterprise-scale identity security programs and govern non-human identities (inventory, ownership, certification, credential rotation, risk).
• Strong cross-functional leadership and communication skills, with sound judgment for operating in an emerging, still-being-defined discipline.
• Must Have: 8+ years in IAM/cybersecurity, with 5+ years at architect or principal level owning target-state design for an enterprise security or identity domain.
• Must Have: Proven experience designing identity, access, and governance controls for non-human identities — service accounts, workload identities, and AI agents — at enterprise scale.
• Must Have: Strong AWS (or equivalent public cloud) identity and access design experience at multi-account scale.
• Must Have: Hands-on experience with Okta (or comparable access management/federation platform) and Saviynt (or comparable IGA platform).
• Must Have: Experience with privileged access management platforms such as Idira (formerly CyberArk) and HashiCorp Vault, or comparable tools.

Qualifications

• Preferred Qualifications: Identity Threat Detection and Response (ITDR) and posture management
• Preferred Qualifications: Experience with SPIFFE/SPIRE or service mesh identity patterns
• Preferred Qualifications: Knowledge of emerging AI regulation and frameworks
• Preferred Qualifications: Externalized authorization and policy orchestration expertise
• Preferred Qualifications: Experience