Security Engineer, Forward Deployed Engineering, AWS Forward Deployed Engineering

About the role

AWS Forward Deployed Engineering (FDE) embeds AI engineers directly inside strategic enterprise customers to design, build, and deploy production-grade AI systems. We are looking for a Security Engineer to join the FDE security team and help secure production AI systems deployed in customer environments.

You will conduct security reviews, write security tooling, perform threat modeling, and implement controls that protect AI systems built by FDE engineers. You work alongside FDE delivery pods in customer environments, getting hands-on with production code, cloud infrastructure, and AI/ML pipelines. You learn fast, ship fast, and build security into the system rather than bolting it on after the fact.

This is a hands-on security engineering role where you spend most of your time writing code, reviewing architectures, and testing systems. You work with senior security engineers who will mentor you, but you own real security deliverables from your first engagement. If you want to learn AI security by doing it, in production, at enterprise scale, this is the role.

This position requires that the candidate selected be a US Citizen.

Key job responsibilities

- Conduct security design reviews, code reviews, and architecture assessments for production AI systems (agentic workflows, RAG pipelines, orchestration layers, model integrations) under guidance from senior security engineers

- Perform threat modeling and penetration testing for AI/ML applications deployed in customer environments; identify vulnerabilities and work with FDE engineers to implement fixes before production release

- Build and maintain security automation: scanning tools, CI/CD security gates, guardrail testing scripts, secrets detection, and dependency vulnerability checks for forward-deployed AI systems

- Implement security controls for AI-specific threats: prompt injection prevention, output filtering, data isolation, model API endpoint hardening, and guardrail configuration

- Support security incident response for FDE-delivered production AI systems; assist with triage, root cause analysis, and remediation under senior engineer guidance

- Document security findings, patterns, and remediation guidance as reusable playbooks and runbooks that benefit future FDE engagements

- Contribute to the FDE security accelerator library: reusable modules, templates, and reference configurations that help FDE engineers build secure AI systems faster

- Requires up to 25% travel

A day in the life

You start the morning running your automated security scan against a new Bedrock-powered application an FDE pod shipped overnight, triaging the findings and filing issues for two real vulnerabilities. Mid-morning you pair with an FDE engineer to fix an IAM policy that's overly permissive for their multi-agent orchestration system. After lunch you're writing a Python script that automates guardrail bypass testing for the team's standard deployment pattern. You close out the day joining a threat modeling session led by a senior security engineer, where you learn how to assess data flow risks in a customer's RAG pipeline. Every week you're working on a different AI architecture, building skills you can't get anywhere else.