AI Security Engineer (Agentic Systems)
ClearRoute14 hours ago
London, United KingdomSenior
Responsibilities
- Design and run red-team exercises and adversarial tests against AI agents, LLMs, and agentic workflows.
- Test for prompt injection, jailbreaking, tool-call and argument manipulation, agent hijacking, memory or context poisoning, and privilege or credential escalation.
- Build repeatable evaluation approaches for non-deterministic systems and define methods to detect, measure, and mitigate hallucinations.
- Own account-wide agentic AI risk identification, governance, testing strategy, and reusable tooling recommendations.
- Map findings and controls to OWASP agent risk categories, NIST AI Risk Management Framework, MITRE ATLAS, and other relevant frameworks.
- Define controls for internal and client-facing agent behavior, including output review, sanitization, and release gating.
- Define rollback and incident-response procedures for compromised or misbehaving agents.
- Track emerging AI regulations and translate them into practical testing and governance requirements.
- Advise delivery teams, represent ClearRoute's agentic security posture, and train security engineers to implement and automate the strategy.
Requirements
- Hands-on experience red-teaming or adversarially testing LLM or agentic AI systems, including prompt injection, jailbreaking, and tool misuse.
- Practical understanding of testing non-deterministic systems and detecting, quantifying, and mitigating hallucinations.
- Working knowledge of agentic-AI risk frameworks and taxonomies such as OWASP agent risk categories and MITRE ATLAS.
- Experience with the OWASP Top 10 for LLM Applications, including prompt injection and system-prompt-leakage risks.
- Experience designing or applying an AI governance framework such as NIST's AI Risk Management Framework.
- Threat modeling and security review experience.
- Ability to operate across accounts and teams and develop testing strategies that generalize beyond one application.
- Strong communication skills for explaining agentic AI risks to technical and non-technical stakeholders.
- Ability to implement and run security tests directly before mentoring and training other security engineers.
- Familiarity with agent-specific evaluation tooling, awareness of AI regulation, multi-agent or agent-mesh systems, cloud security, Azure, Terraform, CI/CD, or regulated-sector environments is preferred.
Benefits
- Genuine hybrid working, real flexibility, a home office budget, and time off including a birthday day off.
- Profit share, recognition programs, private healthcare, competitive salary benchmarking, and an employee assistance program covering counseling, mental health, and financial wellbeing.
- Interest-based communities, monthly socials, cross-regional events, and an inclusive collaborative culture.
- Learning budgets, Udemy access, a clear career framework, internal mobility, and thought leadership opportunities.
- Family-friendly working arrangements and flexibility or interview accommodations for individual needs.
About ClearRoute
ClearRoute is a London‑headquartered platform engineering consultancy that helps enterprises improve software delivery across cloud platforms, quality engineering, and developer experience. Founded in 2022, it provides advisory and hands‑on engineering services, applying its QCE framework to streamline routes to production and manage security, compliance, and scalability in complex, regulated, multi‑vendor environments.