Security Engineer, Security for Employee and Legal Systems (SEALS)

The Security for Employee and Legal Systems (SEALS) team protects Amazon's enterprise HR and Legal applications. Millions of Amazonians use these systems to review their pay, manage their benefits, update their personal information, and handle confidential legal matters. Protecting that data is the core of what SEALS does, and the customers who depend on us are the people who work here. The applications we protect are increasingly built with GenAI, and many of them now embed GenAI features and agentic workflows of their own. That shifts the security work. You will threat model LLM-backed applications, assess risks such as prompt injection, insecure model output handling, and over-broad agent permissions, and secure transitive authentication as agents act on behalf of users. You will also use GenAI to scale your own delivery, from automated code review to detection authoring. SEALS delivers ambient security. Builders get security feedback while they write code, at commit time, so risks are caught and fixed during development. As a Security Engineer on SEALS, you will be an embedded partner to your builder teams and own the security outcomes for your domain. Strong candidates bring depth in application security: threat modeling, architecture and design reviews, secure code review, and hands-on remediation. You will contribute autonomously within the team, deliver risk reduction across the full lifecycle, and mentor peers, while seeking guidance from managers and technical leaders on the hardest tradeoffs. Key job responsibilities ● Own the security outcomes for your assigned SEALS builder teams and the HR and Legal applications they run. ● Threat model GenAI and LLM-backed applications, covering risks such as prompt injection, insecure output handling, data leakage through model context, and excessive agent authority. ● Secure agentic workflows, including transitive authentication and scoped authorization as agents act on behalf of users. ● Author and refine automated detection rules that catch risks at code commit and give builders a fix they can accept immediately. ● Use GenAI to scale delivery across your domain, including AI-assisted code review, detection authoring, and triage. ● Drive paved-path adoption (Secure CDK Blueprints, automated patching, standardized authorization, transitive authentication for agentic workflows) so preventable issues do not occur. ● Hunt for risk the tooling missed through adversarial analysis and manual secure code review in Java, Python, and JavaScript, then turn each escape into a new detection. ● Provide security architecture guidance and lead the security outcomes of design reviews for your team and customers. ● Prioritize remediation by business impact, and escalate to leadership when a builder wants to accept a launch-blocking risk. ● Communicate security outcomes clearly, in writing and verbally, to security and service team leadership. A day in the life You are the security engineer your builder teams want in the room. On a given day you might threat model a new LLM feature for an HR application, write a detection rule that closes a gap a production finding revealed, pair with a builder to fix an authorization flaw while the code is fresh, or review the design of an agentic workflow before it ships. You articulate risk clearly to technical and non-technical audiences, prioritize pragmatically, and guide partners toward secure solutions. You mentor other engineers on the team and raise the bar around you.

SEALS sits within Corporate Services Security (CPSS), the Amazon security team aligned with Finance & Global Business Services, People eXperience & Technology, Legal, and Global Communications and Community Impact. We engage development teams early so our stakeholders can build and scale securely.