Ingénieur de Recherche – Sécurité de l’IA Agentique (H/F)

Develop countermeasures against vulnerabilities in generative models and analyze risks specific to agentic AI systems, including prompt injection and data leakage. Design protection mechanisms such as anomaly detection, adversarial robustness, and homomorphic encryption while evaluating solution efficacy through internal benchmarks and academic collaborations. Contribute to secure system architecture by integrating security into design requirements, collaborating with DevSecOps to automate vulnerability testing, and optimizing solutions for embedded environments. Monitor emerging threat trends, propose risk scenarios to guide R&D roadmaps, and validate solutions in real-world use cases like LLM-based cyberattack detection. Collaborate with product teams to embed security features and train engineers on best practices.