Agentic AI Security Engineer — Red Team & Governance

ClearRoute; Greater London; Hybrid; GBP 85,000 - 120,000; Full time

Benefits offered by this job

Hybrid working

Home office budget

Profit share

Private healthcare

Learning budget

Udemy access

Job summary

ClearRoute is seeking a Senior Security Engineer to own agentic security testing across accounts, designing repeatable evaluation methods and guiding governance for agentic AI. You will work hands-on at first and grow into wider governance and mentorship across teams.

You will map findings to OWASP, MITRE ATLAS and NIST AI RMF, define controls for internal vs client-facing views, and help roll out tooling and incident response processes across accounts.

Qualifications

• Hands-on experience red-teaming or adversarially testing LLM or agentic AI systems.
• Experience testing non-deterministic systems and mitigating hallucinations.
• Familiarity with OWASP/ATLAS risk taxonomies and NIST AI RMF.
• Experience applying an AI governance framework to controls.
• Ability to operate across accounts and teams with test strategy.
• Threat modelling and security review experience.

Responsibilities

• Design and run agentic red-team exercises across AI agents and workflows.
• Develop repeatable evaluation approaches for non-deterministic outputs.
• Define governance, risk identification and tooling recommendations.

Skills

Adversarial testing

Prompt injection

Non-deterministic testing

OWASP/MITRE

AI governance

Cross-account work

Threat modelling

Communication

Tools

Terraform

Azure

CI/CD

Job description

ClearRoute is seeking a Senior Security Engineer to own agentic security testing across accounts, designing repeatable evaluation methods and guiding governance for agentic AI. You will work hands-on at first and grow into wider governance and mentorship across teams.

You will map findings to OWASP, MITRE ATLAS and NIST AI RMF, define controls for internal vs client-facing views, and help roll out tooling and incident response processes across accounts.