Agentic AI Security Engineer: Sandbox & Guardrails

Howard Hughes Medical Institute (HHMI); Ashburn, Northern (VA, KY); Hybrid; USD 149,000 - 186,000; Full time

Benefits offered by this job

Health benefits

Job summary

HHMI's Janelia Research Campus seeks an Agentic AI Security Engineer to design and operate the safety stack for agentic AI systems. You will implement sandboxing, guardrails, and real-time observability to ensure secure AI agent behavior and auditable sessions.

The role emphasizes hands-on security engineering, integration with OpenShift/Kubernetes environments, and collaboration across biology, ML, and tooling teams.

Qualifications

• Minimum 5 years in security, systems/SRE, or ML systems engineering.
• Hands-on experience with LLM agents and sandbox/container-security.
• Experience with Linux and Kubernetes/OpenShift isolation and governance.

Responsibilities

• Research, develop, deploy, operate, and analyze the safety stack for agentic AI at Janelia.
• Implement sandboxing, guardrails, and observability across agent sessions and tool calls.
• Collaborate on architecture decisions and publish or open-source work where appropriate.

Skills

LLM agents

Sandboxing

Kubernetes/OpenShift

Agent architectures

Open-source contributions

Security/ML systems (5+ yrs)

Tools

MCP protocol

Job description

HHMI's Janelia Research Campus seeks an Agentic AI Security Engineer to design and operate the safety stack for agentic AI systems. You will implement sandboxing, guardrails, and real-time observability to ensure secure AI agent behavior and auditable sessions.

The role emphasizes hands-on security engineering, integration with OpenShift/Kubernetes environments, and collaboration across biology, ML, and tooling teams.