AI Security Engineer

About the mandate

You secure AI solutions, models, agents, data and integrations at a large public sector organisation in Brussels. You also help deploy AI responsibly for detection, analysis, vulnerability management, patch prioritisation and secure development, with human responsibility and oversight remaining the starting point. You carry out threat modelling and security assessments, develop guardrails, test misuse scenarios and translate findings into actionable measures. You work with colleagues from IT, business, security and architecture, and explain technical concepts clearly to both technical and non-technical stakeholders. The assignment runs for 12 months from 14/12/2026 and you can work remotely part-time.

Responsibilities

• Carry out threat modelling and risk and security assessments for GenAI, ML and agentic AI solutions
• Develop guardrails for identity, least privilege, secrets, network isolation, data and prompt security, tool usage, output validation, logging and monitoring
• Test misuse scenarios, such as prompt injection, unwanted data exfiltration, excessive permissions, unreliable output and misuse of tools or agents
• Integrate security requirements into sandboxes, gateways, pipelines, model and dependency management and production platforms
• Help design security monitoring and incident response for AI components, including ownership, audit trails and escalation
• Assess suppliers, models, libraries and external AI services as part of supply chain risk

Must-haves

• Demonstrable experience in cybersecurity engineering, cloud security, application security, IAM, SOC or DevSecOps
• Understanding of the specific risks of LLMs, agents, retrieval and model and data supply chains
• Experience with security testing, technical risk assessments and translating findings into actionable measures
• Ability to enable innovation with proportionate controls, without reducing security to a checklist
• Knowledge of GDPR, AI Act, NIS2 or similar frameworks
• You follow new trends and experiment with innovative AI technologies
• Strong in analysing, advising, structuring and guiding multidisciplinary decision-making
• Good communication with technical and non-technical stakeholders and keeping to deadlines

Nice-to-haves

• Knowledge of the public sector, G-Cloud or the OISZ context

What the employer offers

• Place of work in Brussels
• Possibility to work remotely part-time
• 12-month assignment