AI Information Security Engineer

Southern New Hampshire University is a team of innovators committed to transforming lives through education. They are seeking an AI Information Security Engineer to secure AI systems and models by implementing security measures and collaborating with various teams to ensure safety and compliance in production environments.

Responsibilities:

• Document AI system components and data flows, including prompts, context, embeddings, training data, model artifacts, outputs, and agent tool interactions
• In collaboration with the AI team, identify attack surfaces, trust boundaries, and privilege transitions within AI pipelines and agent workflows and perform structured threat modeling for AI systems during design, development, and change cycles in collaboration with the AI team
• In collaboration with the AI team, translate identified threats into concrete, relevant security requirements and engineering tasks in collaboration with the AI team
• Implement technical controls informed by established AI security frameworks (e.g., OWASP LLM Top 10, NIST AI RMF) according to compliance requirements and AI governance guidance
• Design, build, and maintain automated security testing for AI systems within CI/CD pipelines, supports testing for prompt injection, unsafe model behavior, misconfigured access, data exposure, and agent misuse. Ensure AI security controls are validated during build, deployment, and change cycles, with failures surfaced early to engineering teams
• Implement technical guardrails to protect sensitive data used by AI systems, including retrieval of augmented generation (RAG) pipelines and external data sources
• In collaboration with the AI Team, Design and operate controls for sensitive data identification, minimization, redaction, and leakage prevention—addressing PII and other protected data in prompts, context, embeddings, and outputs to ensure privacy preserving AI operation in production environments
• Design, implement, and maintain security controls across the full AI/ML lifecycle—including data ingestion, training, evaluation, deployment, inference, and CI/CD—covering model artifacts, configurations, embeddings, prompts, and deployment patterns
• Implement and operate runtime safeguards for AI services and agent-based systems, including input and output controls, context isolation, tool use restrictions, and abuse prevention mechanisms (e.g., rate limiting and anomaly detection), ensuring safe operation without breaking functional requirements
• Design security controls that balance safety, system performance, reliability, and developer usability in production of AI services
• Implement and operate secure identity, secrets, and access control patterns for AI services, agents, and integrations, enforcing least privilege, integrating with enterprise IAM and key management systems, and monitoring credential usage and rotation
• Instrument AI systems to produce actionable logging, metrics, and traces; build dashboards and alerts for detecting prompt manipulation, anomalous usage, and unexpected behavior; and integrate AI specific signals into enterprise security operations workflows
• Embed with AI engineering and platform teams to design and maintain technical security controls; develop reusable security components and patterns; contribute documentation and runbooks; and, in collaboration with the AI team, communicate AI security requirements and remediation outcomes to technical, non-technical, and cross functional stakeholders

Requirements:

• 5+ years of experience in IT or cybersecurity, with engineering responsibilities (i.e. IT Security or Application Development)
• 2 + years of experience securing AI/ML systems or adjacent domains with demonstrated application to AI workloads
• Experience with security engineering principles, including authentication, authorization, logging, and monitoring
• Experience with AI/ML concepts such as models, training data, inference pipelines, embeddings, and agent frameworks
• Experience modeling data flows, trust boundaries, and attack paths in AI systems
• Experience mitigating threats such as prompt injection, model poisoning, model theft, and data leakage
• Experience implementing controls such as input validation, output filtering, context isolation, and abuse detection