Cybersecurite à France

Organisation/Company Grenoble INP - Institute of Engineering Department Engineering Research Field Engineering Researcher Profile Other Profession Positions PhD Positions Country France Application Deadline 4 Jan 2026 - 20:00 (Europe/Paris) Type of Contract Temporary Job Status Full-time Offer Starting Date 1 Feb 2026 Is the job funded through the EU Research Framework Programme? Not funded by a EU programme Is the Job related to staff position within a Research Infrastructure? No

Project summary:

The growing integration of Artificial Intelligence (AI) modules into safety-critical embedded systems (autonomous vehicles, drones, industrial and medical devices) raises major safety and security concerns. These modules, often based on deep neural networks, are sensitive to both accidental faults and intentional attacks [1]-[5] that can alter their decisions. Ensuring their robustness under real-world conditions is therefore essential for trustworthy deployment.

Current approaches mainly focus on software-level adversarial robustness or high-level fault tolerance. However, few methodologies jointly consider physical disturbances, embedded constraints, and AI decision integrity in critical systems.

Main goals:
This PhD aims at developing a unified methodology for evaluating and improving the robustness of embedded AI modules against various real-world and physical disturbances, encompassing both safety-related faults and security-related attacks.

The work will: (1) Identify, model, and reproduce representative perturbations that may cause abnormal or unsafe behavior. (2) Evaluate their effects on performance, safety, and security metrics. (3) Propose and validate mitigation and hardening techniques at the model, system, and learning levels.

The targeted application will concern multi-sensors based systems for autonomous vehicles embedding AI perception or decision modules.

The experiments will rely on embedded platforms available at LCIS, including electromagnetic fault injection benches. The methodology will address both safety-related disturbances (accidental faults) and security-related threats (intentional perturbations), highlighting the differences between them in embedded AI systems.

Keywords : Embedded AI, robustness, fault injection, safety, security, adversarial attacks, physical attacks, autonomous systems, multi-sensor systems, hardening techniques, critical embedded systems

E-mail job-ref-7p6x515qyh@emploi.beetween.com

Research Field Engineering Education Level Master Degree or equivalent

Skills/Qualifications

Core competencies:

• Prototyping and Simulation of Digital Systems