AI Security & Compliance Engineer

Role purpose

Ensure AI and GenAI systems on AIRP are designed, deployed, and operated securely and in compliance with enterprise technology, cybersecurity, privacy, and regulatory standards. The role covers emerging LLM risks as well as traditional AWS cloud, application, data-security, DevSecOps, and IaC controls.

Primary ownership

Security architecture and control implementation for AI platforms, LLM applications, RAG pipelines, model-serving environments, and agentic systems.

Threat modeling, AI red teaming, vulnerability assessment, risk remediation, and secure production approvals.

Security evidence, control documentation, and compliance support for AIRP releases, Terraform/IaC, and DevOps pipelines.

Key responsibilities

Design and review secure architectures for AI/ML platforms, LLM applications, RAG pipelines, model-serving environments, and agentic AI workflows.

Conduct threat modeling for prompt injection, jailbreaks, insecure tool use, model inversion, data leakage, retrieval poisoning, adversarial inputs, unauthorized access, and third-party model risk.

Implement controls for AWS IAM, encryption, key management, secrets management, network segmentation, API security, logging, secure data handling, and data-loss prevention.

Embed security into MLOps, LLMOps, CI/CD, container security, infrastructure-as-code, Terraform modules, and deployment pipelines.

Review cloud-agnostic IaC templates and AWS-specific deployments for least privilege, secure defaults, segregation of duties, policy compliance, and auditability.

Must-have candidate profile

Strong background in cybersecurity, cloud security, application security, DevSecOps, or technology risk.

Experience securing cloud-native platforms, APIs, microservices, containers, Kubernetes, CI/CD pipelines, and infrastructure-as-code.

Strong AWS cloud security exposure or comparable hyperscaler security depth, including IAM, encryption, network controls, logging, secrets, and secure deployment patterns.

Understanding of AI/ML and GenAI-specific risks such as prompt injection, adversarial attacks, data leakage, model misuse, retrieval poisoning, model supply-chain risk, and unsafe tool use.