NAVA Software Solutions is seeking an AI Security Engineer to work on AI model security for their client TIAA. The role involves designing and implementing security controls for AI systems, conducting vulnerability assessments, and establishing governance standards.
Responsibilities:
- Design, implement, and maintain AI safety and security controls for LLM-based and agentic AI systems, including prompt injection defenses, jailbreak prevention, PII protection, and tool-use restrictions
- Conduct AI red teaming and adversarial testing to identify vulnerabilities, misuse scenarios, prompt attacks, data leakage risks, and unauthorized agent behaviors
- Develop and maintain automated safety evaluation frameworks and regression test suites to validate AI systems against security, compliance, and responsible AI requirements before deployment
- Perform threat modeling and risk assessments for agentic AI applications, identifying potential attack vectors related to memory, retrieval systems, tool calling, and external integrations
- Implement runtime monitoring, detection, and incident response capabilities to identify anomalous agent behavior, policy violations, excessive tool usage, and potential security incidents
- Establish and enforce AI governance, security, and responsible AI standards, translating organizational policies and regulatory requirements into technical controls and measurable acceptance criteria
Requirements:
- AI Model Security Threat scanning
- Agentic AI foundational security controls
- Agentic AI security development Lifecycle