Agentic AI Security Specialist

Vbeyond; Greater London; On-site; GBP 120,000 - 150,000; Full time

Job summary

Vbeyond in the United Kingdom is seeking a seasoned security leader to oversee Agentic AI, autonomous workflows, and multi-agent systems. You will protect agents, tools, memory, and data from misuse, adversarial attacks and unintended actions, while shaping governance and secure execution.

The role requires deep hands-on expertise in AI agent orchestration, red teaming, threat modelling and enterprise-grade security controls, with collaboration across platforms and enterprise teams.

Qualifications

• Minimum 8 years of relevant professional experience, including substantial hands-on experience in the core specialisms of this role.
• Deep understanding of Agentic AI and LLM security.
• Hands-on experience with AI agent orchestration, RAG, tool calling, APIs and multi-agent architectures.
• Strong experience in AI red teaming, adversarial testing and threat modelling.
• Understanding of AI governance, security guardrails, responsible AI and model risk.
• Strong cloud, application and identity security expertise.
• Ability to translate emerging AI security risks into enterprise-grade security controls.
• Strong understanding of agent identity, authorisation, secrets management, auditability and human oversight.

Responsibilities

• Lead threat modelling for AI agents, tool-calling architectures, autonomous workflows and multi-agent ecosystems.
• Assess security risks across prompts, models, memory, retrieval, APIs, plugins, tools and identity boundaries.
• Conduct AI red teaming and adversarial testing for agent manipulation, excessive agency, prompt injection and data leakage.
• Design security guardrails, approval controls, isolation patterns, policy enforcement and secure execution boundaries.
• Define security requirements for agent identity, authorisation, secrets, auditability and human oversight.
• Partner with platform and application teams to securely integrate agents with enterprise systems.
• Support monitoring, incident response and risk mitigation for Agentic AI solutions.
• Track emerging threats, standards and security control patterns relating to autonomous AI.

Skills

Agentic AI

LLM security

Threat modelling

Red teaming

Governance & risk

Cloud security

Identity security

Auditing & oversight

Education

Bachelor's in Cyber Security

Equivalent professional experience

Job description

Role Purpose

Provide specialist security leadership for Agentic AI, autonomous workflows, multi-agent systems and LLM-powered applications. The role will focus on protecting AI agents, tools, memory, data and orchestration layers from misuse, adversarial attacks and unintended actions.

Key Responsibilities

• Lead threat modelling for AI agents, tool-calling architectures, autonomous workflows and multi-agent ecosystems.
• Assess security risks across prompts, models, memory, retrieval, APIs, plugins, tools and identity boundaries.
• Conduct AI red teaming and adversarial testing for agent manipulation, excessive agency, prompt injection and data leakage.
• Design security guardrails, approval controls, isolation patterns, policy enforcement and secure execution boundaries.
• Define security requirements for agent identity, authorisation, secrets, auditability and human oversight.
• Partner with platform and application teams to securely integrate agents with enterprise systems.
• Support monitoring, incident response and risk mitigation for Agentic AI solutions.
• Track emerging threats, standards and security control patterns relating to autonomous AI.

Required Skills and Experience

Minimum 8 years of relevant professional experience , including substantial hands-on experience in the core specialisms of this role.

• Deep understanding of Agentic AI and LLM security .
• Hands-on experience with AI agent orchestration, RAG, tool calling, APIs and multi-agent architectures .
• Strong experience in AI red teaming, adversarial testing and threat modelling .
• Understanding of AI governance, security guardrails, responsible AI and model risk .
• Strong cloud, application and identity security expertise.
• Ability to translate emerging AI security risks into enterprise-grade security controls .
• Strong understanding of agent identity, authorisation, secrets management, auditability and human oversight .

Qualifications and Certifications

• Bachelor's degree in Cyber Security, Computer Science, Engineering, Data Science or a related discipline, or equivalent professional experience.
• CISSP, CCSP, CISM or equivalent certification is desirable.
• OSCP, CEH, GIAC or equivalent offensive security certification is advantageous.
• AI governance, Responsible AI or ML security certification is desirable.

Core Behaviours

• Strong ownership, analytical thinking and delivery focus.
• Ability to collaborate across UK and offshore multidisciplinary teams.
• Clear communication with technical and non-technical stakeholders.
• Commitment to secure, ethical and compliant engineering practices.
• Curiosity and continuous learning in a rapidly evolving AI security landscape.