Agentic AI Security Engineer
Core
Research, develop, deploy, and operate the safety stack for agentic AI systems, focusing on sandboxing, guardrails, and observability to prevent hallucinated commands and runaway loops.
Role type
Senior IC security engineer specializing in agentic AI safety
Builds
Hardened execution environments, policy enforcement guardrails, and real-time monitoring systems for AI agents
Domain
AI Safety / Systems Security / Research Infrastructure
Deliverable
production ML models | infrastructure
Required skills
LLM agent operations, container and kernel-level sandboxing, Kubernetes/OpenShift isolation, agent architecture protocols (MCP), adversarial testing, audit logging
Preferred skills
Open-source contributions, technical publications, disclosed security findings
Technologies
NVIDIA GPUs (B300/H200/H100), OpenShift, Kubernetes, Linux, MCP, judge models
Responsibilities
Design and implement sandboxing environments for agent code; enforce policy guardrails on agent inputs and outputs; build observability tools for agent session tracing and incident analysis
Seniority
Senior, hands-on IC