AI Security Specialist

Zerolynx logoZerolynxMadrid, Spain (Hybrid) · Full-time · Hybrid

Founded 2018 · HQ Madrid, Spain · 100+ employees

About this role

Zerolynx is a Spanish cybersecurity company based in Madrid. It is looking for an AI Security Specialist to join client projects in a hybrid setup, helping teams build AI agents safely.

The role

Support the client's team in the design and creation of their AI agents, analyze the security of the AI components involved in decision-making, and define the MCP connectors that link services with documentation from various sources. Your responsibilities will be:

AI agent design

  • Support the client's team in the design and creation of their AI agents.
  • Assist them in determining which use cases are suitable for automation, identifying the tools and data each agent should access, defining where human intervention is required, and establishing an overall governance framework.
  • Review projects and provide best practices.

AI security analysis

  • Analyze the security of the AI components involved in decision-making.
  • Identify risks such as prompt injection, data leakage, manipulation of the information the model relies on, and excessive agent permissions.
  • Verify that decisions are traceable and explainable.

MCP connectors

  • Define the MCP (Model Context Protocol) connectors that will link ASO services with documentation from various sources.
  • Specify what each connector should expose, including the necessary permissions and controls, and validate that the final implementation meets these requirements.

Requirements

  • Practical knowledge of generative AI (ideally with corporate experience), specifically regarding agents, assistants, or RAG (Retrieval-Augmented Generation) systems. A solid understanding of their internal mechanics is required, even if you are not directly involved in programming them.
  • Familiarity with MCP and how agents integrate with other systems and documentation sources.
  • A strong foundation in AI security, such as the OWASP Top 10 for LLMs and agentic applications, or the MITRE ATLAS framework, along with experience conducting risk assessments or security reviews of AI-based systems.
  • Consulting skills: the ability to work closely with the client's team, explain risks to both technical and business stakeholders, and ensure thorough documentation.

Nice to have

  • Knowledge of the regulatory landscape, including the EU AI Act, ISO/IEC 42001, and the NIST AI RMF.
  • Experience participating in AI system audits.
  • Experience in AI governance, such as maintaining model inventories, establishing usage policies, and conducting risk assessments.
  • Experience working with cloud-based AI platforms, such as Azure OpenAI, AWS Bedrock, or Google Vertex.