AI Safety Research Scientist

Responsibilities

• Help clarify which AI safety research questions are central to the Scientist AI agenda
• Analyze proposed solutions to research questions and develop novel solution ideas
• Propose, design, and implement novel ML algorithms and experiments to solve or test complex AI safety problems
• Scrutinize the Scientist AI agenda and its strategic fit for mitigating AI risk to guide and prioritize experiments

Requirements

• Research experience in technical AI safety, ideally with a focus on AGI safety
• Experience proposing, designing, and running ML experiments
• Proficiency with ML frameworks such as PyTorch or TensorFlow
• Ability to work in a collaborative environment and discuss complex findings with diverse technical teams
• PhD in a relevant field preferred (equivalent experience accepted)

Preferred Qualifications

• Experience in AI safety research areas: agent foundations, AI control, frontier AI safety, eliciting latent knowledge, externalised reasoning, or ambitious alignment schemes
• Machine learning research publications in top-tier venues (e.g., NeurIPS, ICML, ICLR)
• Experience fine tuning and post training language models

Benefits

• Comprehensive health benefits
• Minimum of 20 days vacation per year
• Minimum of 4% retirement savings employer contribution
• Generous flexible benefits
• Collaborative and inclusive work environment

About the Company

LawZero is a non-profit organization committed to advancing research and creating technical solutions that enable safe-by-design AI systems. Based in Montreal, LawZero’s research aims to build non-agentic AI that learns primarily to understand the world rather than to act in it, providing truthful answers through transparent and externalized probabilistic reasoning.