Amazon’s Security team is seeking an AI Security Engineer to help protect AI applications used by large numbers of customers. This onsite role in Jessup, MD focuses on partnering with software development teams across threat modeling, secure code review, AI security research, automation, training, and architecture guidance.
Role Focus
The AI Security Engineer will improve the security posture of AI systems by creating and maintaining threat models, conducting manual and automated code reviews, and developing AI security automation tools. The role also emphasizes independent problem solving for AI security issues that require novel methods, along with guidance to influence engineering process and priorities to improve outcomes.
Key Responsibilities
- Create, update, and maintain threat models across a variety of software projects.
- Perform manual and automated code review, with primary focus on Java, Python, and JavaScript.
- Develop AI security automation tools.
- Conduct AI security training and outreach for internal development teams.
- Provide AI security architecture and design guidance.
- Independently address AI security problems that require new or unconventional approaches.
- Drive improvements by influencing team and partner process, priorities, and decision-making.
Required Qualifications
- Bachelor’s degree in a STEM field such as Computer Science, Computer Systems Engineering, Software Development, Mathematics, or a similar discipline.
- 3+ years of any combination of: threat modeling, secure coding, identity management and authentication, software development, cryptography, system administration, and/or network security.
- Programming experience in Python, Ruby, Go, Swift, Java, .Net, C++, or a similar object-oriented language.
- Knowledge of machine learning concepts and the ML model lifecycle (training, deployment, inference), plus familiarity with AI/ML-specific threats such as adversarial inputs, model evasion, data poisoning, and model extraction.
- Current, active US Government Security Clearance of TS/SCI with Polygraph.
Technologies
- Java
- Python
- JavaScript
- Ruby
- Go
- Swift
- .Net
- C++
Preferred Qualifications
- Knowledge of networking fundamentals, security, and databases (relational or NoSQL), along with operating system fundamentals (Unix, Linux, and/or Windows).
- Hands-on experience securing or attacking AI/ML systems, including adversarial ML, red-teaming models, or ML threat detection, with familiarity such as MITRE ATLAS.
- Security certification including one or more of: GCIH, GSEC, GCFA, GREM, Security+, CISSP, CISA, or CISM.
- Experience configuring security automation (static/dynamic analysis, code-quality, container scanning) and integrating AI/ML security tooling into pipelines.
Compensation and Location
Location: Jessup, MD (onsite).
Salary: USD 159,300 - 202,400 per year.
Benefits
- Sign-on payments
- Restricted stock units (RSUs)
- Health insurance (medical, dental, vision, prescription)
- Basic Life & AD&D insurance
- Option for Supplemental life plans
- EAP
- Mental Health Support
- Medical Advice Line
- Flexible Spending Accounts
- Adoption and Surrogacy Reimbursement coverage
- 401(k) matching
- Paid time off
- Parental leave
A Day in the Life
The day begins with a review of model telemetry, focusing on anomalies surfaced overnight. The engineer triages findings, then initiates investigations into adversarial input patterns against deployed models. Forensic data is gathered to understand what occurred, the incident is driven through containment, and the learning is translated into new detections.
Later, the engineer reviews scanning rules and integrates an AI/ML security automation tool into the pipeline. The afternoon includes documenting after-action findings and coordinating reporting across the team.
