Lensys MarketplaceMissions
Lensys
Suivez-nous

AI Security Engineer

Protect an ambitious AI estate by embedding security into models, agents, data flows, and platforms. This mission blends threat modelling, guardrail engineering, abuse testing, and incident response for a 12-month hybrid engagement in Brussels.

CONTEXT

Our client, a major player in its sector, is expanding its use of AI across services where trust, privacy, and continuity are non-negotiable. The organisation is not simply adding security as a final checkpoint; it wants security designed into the entire AI lifecycle, from data preparation and model selection to agent behaviour, platform integration, and incident response. The AI Security Engineer will act as a bridge between AI builders and security teams, turning abstract risks into concrete guardrails, testable abuse scenarios, and operational playbooks that keep intelligent systems useful without becoming unpredictable or exposed.

ACTIVITIES

  • Perform threat modelling for AI solutions, models, agents, and data flows across the full lifecycle.
  • Design and implement security guardrails, controls, and monitoring patterns for AI platforms and applications.
  • Develop and execute abuse-case testing, red-team scenarios, and security validation for agentic and model-driven systems.
  • Integrate security requirements into AI platforms, CI/CD pipelines, and engineering standards.
  • Contribute to incident response design, detection playbooks, and cross-team escalation paths for AI-related security events.
  • Advise product and engineering teams on secure AI patterns, data protection, model access, and safe deployment practices.

EXPECTED ACHIEVEMENTS

  • A practical AI security framework embedded into platforms, delivery processes, and day-to-day engineering.
  • A library of tested guardrails and abuse scenarios that reduce exposure before systems reach production.
  • Clear incident response playbooks and improved readiness for AI-related security events.
  • Stronger collaboration between AI builders, security teams, and business owners, with security becoming an enabler rather than a late-stage blocker.

ENVIRONMENT

  • AI/LLM platforms, agent frameworks, model APIs, vector stores, and data pipelines.
  • Security engineering practices including threat modelling, guardrail design, abuse testing, and incident response.
  • Cloud-native architectures, CI/CD, logging, monitoring, and detection tooling.
  • Agile delivery with cross-functional collaboration across security, data, and engineering teams.