Principal AI Security Engineer

EY; Chantilly (VA); Remote; USD 125,000 - 230,000; Full time

Job summary

EY seeks an AI Security Engineer to own the security posture of the agentic AI platform end to end, spanning from silicon-level attestation to prompt-injection defenses. The role requires deep cloud-native security expertise and hands-on work with AI threat models and defense-in-depth across the stack.

You will lead security reviews for regulated deployments, drive security by design, and coordinate with clients and regulators to provide assurance artifacts that unblock deployment into sensitive

Qualifications

• Bachelor’s or Master’s degree in Computer Science, Security, or a related technical field.
• 10+ years in security engineering or offensive security with hands-on ownership.
• Hands-on production security in cloud-native environments.

Responsibilities

• Own the platform threat model across agent autonomy, tool invocation, and data supply chains.
• Define security controls for infrastructure, Kubernetes, identity, and data protection.
• Lead policy-as-code with Kyverno and OPA; enforce guardrails in CI/CD and runtimes.
• Lead AI red teaming and adversarial testing before client/regulator reviews.
• Drive supply-chain integrity: signing, SBOM, provenance, and attestation.

Skills

Cloud-native security

AI threat models

Zero-trust architecture

Policy-as-code

Software supply-chain security

Red teaming / adversarial testing

Executive communication

Education

Bachelor’s or Master’s in CS or Security

Tools

Sigstore/Cosign

SPIFFE/SPIRE

Vault/KMS or equivalent

Job description

EY seeks an AI Security Engineer to own the security posture of the agentic AI platform end to end, spanning from silicon-level attestation to prompt-injection defenses. The role requires deep cloud-native security expertise and hands-on work with AI threat models and defense-in-depth across the stack.

You will lead security reviews for regulated deployments, drive security by design, and coordinate with clients and regulators to provide assurance artifacts that unblock deployment into sensitive