AI Information Security Engineer

Southern New Hampshire University

2d•$94,130 - $150,634•Remote

About The Position

Southern New Hampshire University (SNHU) is seeking an AI Security Engineer to join their innovative team. This is a hands-on security engineering role focused on securing AI systems, models, and agent-based workloads throughout their development lifecycle. The engineer will be responsible for threat modeling, implementing controls, testing, and monitoring to protect AI training data, inference pipelines, agents, tools, and generated outputs from misuse or compromise. The role involves partnering with AI engineering, platform, and security teams to ensure AI systems are secure by design, resilient, and observable in production. The position is 100% remote and requires residency in one of the approved states.

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.

Responsibilities

Document AI system components and data flows, including prompts, context, embeddings, training data, model artifacts, outputs, and agent tool interactions.
Identify attack surfaces, trust boundaries, and privilege transitions within AI pipelines and agent workflows.
Perform structured threat modeling for AI systems during design, development, and change cycles.
Translate identified threats into concrete, relevant security requirements and engineering tasks.
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, supporting 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.
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 communicate AI security requirements and remediation outcomes to technical, non-technical, and cross functional stakeholders.