Staff AI Security Specialist

About the role

ServiceNow is looking for a Staff AI Security Specialist who sits at the intersection of AI and security and enjoys building. The team secures the AI in ServiceNow's products and platform from adversaries, and the role is applied: you reason about how agentic systems break and write the code that stops them, from prototype to shipped control, in collaboration with Security Engineers.

Key responsibilities

• Analyze the security of AI agent systems, including how they assemble context, call tools, plan and delegate, and respond to untrusted input.
• Design and build security controls: static analysis of agent and tool configurations, runtime policy enforcement, behavioral detection, and mitigations for tool and goal hijacking.
• Build working prototypes end to end, then partner with engineering to harden the ones that prove out.
• Threat model new AI architectures and features before they ship, turning findings into concrete controls.
• Red-team agents and defenses against indirect prompt injection, tool abuse, privilege escalation, and data exfiltration.
• Build repeatable evaluations and benchmarks to measure whether controls hold.
• Track offensive and defensive AI security research and advise product and platform teams on secure design.
• Publish internal research, patents and external talks.

Required profile

• Experience integrating AI capabilities into work processes, decision-making or problem-solving.
• 6+ years combined in security, software engineering or applied research, including 2+ years hands-on with AI/ML or LLM-based systems.
• Demonstrated depth in at least one security domain (application security, offensive security, authorization and identity, or detection engineering).
• Ability to take an ambiguous problem to a working artifact and to explain technical concepts to business users and technical stakeholders.

Required skills

• Strong working knowledge of the agent harness: system prompts, context windows, tool-calling loops, sub-agent delegation, sandboxing and execution boundaries.
• Knowledge of agent memory and retrieval systems, orchestration frameworks and tool protocols such as MCP.
• Understanding of AI threats: prompt, tool and goal hijacking, memory poisoning, OWASP Top 10 for LLM applications, MITRE ATLAS.
• Proficiency in Python, enough to build and ship systems independently.
• Preferred: runtime security tooling, static/dynamic analysis, AI red-teaming, published security or AI research.