Chercheur·euse en sciences mathématiques – Sécurité de l'IA

About the role

LoiZéro, a non‑profit founded by Yoshua Bengio, is building AI systems that are safe by design. We are seeking an exceptional theoretical researcher in mathematical sciences to advance our AI safety agenda, working closely with world‑class scientists and engineers.

Key responsibilities

• Explore novel AI system architectures, theoretical frameworks, and safety mechanisms to define a safe strong AI (AGI) model and its guarantees.
• Collaborate with Yoshua Bengio and leading researchers to formalize and refine safety requirements, including probabilistic guarantees under finite data and compute constraints.
• Analyze, extend, and publish the existing theoretical work at LoiZéro, contributing to scientific papers and conferences.
• Maintain rigorous scientific monitoring of AI safety and machine‑learning research, understanding risks such as misalignment, reward hacking, and instrumental goals.
• Help shape and prioritize the overall research programme, ensuring critical safety questions are addressed without hindering capability progress.
• Share AI‑safety expertise with engineers and researchers to improve algorithm design and experimental implementation.

Required profile

• PhD in machine learning, computer science, mathematics, statistics, or a closely related field.
• Demonstrated ability to conduct high‑impact theoretical research and publish in top venues.
• Strong analytical skills and a passion for tackling fundamental AI safety challenges.

Required skills

• Deep knowledge of mathematical foundations of machine learning.
• Experience with formal methods, probabilistic reasoning, and theoretical analysis.