Senior AI Security Engineer: Build Safe AI Systems

EY; San Mateo (CA); Hybrid; USD 134,000 - 201,000; Full time

Job summary

EY is seeking a security-focused leader in the US to embed safety into AI across data, tools and business processes. You will partner with product and engineering teams to design reusable security patterns, threat models, and controls that scale and prove resilience under adversarial conditions.

The role emphasizes hands-on production security, threat modeling, and cross-functional collaboration to deliver measurable protection while enabling rapid innovation. Hybrid travel up to 50% is expected.

Qualifications

• 5+ years of relevant experience.
• Bachelor’s Degree in a relevant field.
• Substantial experience in product, application, cloud, platform, or detection security, including hands-on ownership of production controls.
• Threat modeling: assess complex systems and turn attack paths into practical engineering measures.
• Production security: securing modern apps, cloud environments, identities, data, and pipelines.
• AI systems knowledge: understanding of generative and agentic AI, including retrieval, model access, tools, evaluation, and runtime behavior.
• Hands-on automation: building or extending security tests, controls, telemetry, or reference implementations.
• Response experience: supporting vulnerability or incident response and translating lessons into stronger controls.
• Cross-functional delivery: partnering with engineers and risk functions while balancing user value.
• Travel and onsite: up to 50% travel; hybrid office.

Responsibilities

• Set the security direction. Own reusable AI security patterns, control objectives, reference architectures, and threat models aligned to the risk and autonomy of each use case.
• Join builders early. Work with teams before key design choices are fixed, identifying credible attack paths and shaping controls that fit the product and delivery model.
• Design scoped authority. Create practical patterns for identity, access, isolation, approvals, and recovery so tool-using AI acts within clear and enforceable boundaries.
• Protect sensitive information. Embed security across the data lifecycle, preserving confidentiality, integrity, provenance, permitted use, and monitored access.
• Secure the delivery path. Integrate automated security tests, policy gates, trusted artifacts, traceable changes, and safe rollback into engineering workflows.
• Test adversarially. Run focused red-team exercises against AI-specific abuse cases and convert findings into repeatable tests, stronger platform patterns, and actionable guidance.
• Build runtime protection. Use end-to-end evidence and behavioral signals to detect abnormal actions, privilege misuse, attempted exfiltration, and degrading controls.
• Strengthen response. Support vulnerability triage and incidents with clear evidence, containment options, recovery actions, and lessons that improve future designs.
• Advise on residual risk. Give product, security, evaluation, and risk leaders concise technical recommendations while keeping approval and risk‑acceptance decisions with accountable owners.
• Build shared capability. Create reference implementations, coach engineers, and turn production lessons into reusable controls that make the safer path easier to adopt.

Skills

Security engineering depth

AI threat judgment

Control engineering

Adversarial mindset

Automation capability

Pragmatic influence

Education

Bachelor’s Degree in a relevant field

Job description

EY is seeking a security-focused leader in the US to embed safety into AI across data, tools and business processes. You will partner with product and engineering teams to design reusable security patterns, threat models, and controls that scale and prove resilience under adversarial conditions.

The role emphasizes hands-on production security, threat modeling, and cross-functional collaboration to deliver measurable protection while enabling rapid innovation. Hybrid travel up to 50% is expected.