Apple
AI Security Architect - ARC, AIS
$176K–$264K
Cupertino, California · Full-time · Posted Sep 28, 2026
- Experience
- 6+ years of equivalent, hands-on security experience in large enterprise environments is stated as an alternative to a degree; the posting also describes this experience as a plus.
- Education
- BS in Computer Science, Computer Engineering, or Information Security, or 6+ years of equivalent, hands-on security experience in large enterprise environments.
About the job
Shapes security architecture and reusable safeguards for high-risk AI initiatives, while applying AI tools to scale security reviews and guidance across Apple.
What you'll do
- Conduct AI security architecture reviews, threat modeling, and risk assessments; select or build security solutions.
- Create reusable secure-by-default patterns and AI-powered guidance for engineering and product teams.
- Apply AI-driven tooling to automate threat modeling, design reviews, and control validation.
What they ask for
- Strong background in security architecture, application security, or cloud security.
- Understanding of AI/ML concepts, pipelines, and risks, plus experience with threat modeling and secure design principles.
- Familiarity with AI regulatory frameworks; experience using AI/LLMs for security automation.
- Proficiency in at least one scripting or programming language and ability to communicate security concepts clearly.
Nice to have
- Security architecture reviews spanning cloud-native and AI/ML technologies; security patterns for LLM-based systems, RAG pipelines, or agentic architectures.
- Familiarity with Zero Trust and Secure by Default; data protection, encryption architecture, and key management in AI/ML contexts.
Pay
$175,500 to $263,800 USD base salary
Tools
AI/ML, STRIDE, EU AI Act, NIST AI RMF, ISO/IEC 42001, Python, Swift, Java
AI focus
Secures AI/ML systems against risks such as data poisoning and prompt injection, and applies AI/LLMs to automate security workflows and deliver scalable guidance.
Model-written summary · the original posting is authoritative