About the job
We are looking to hire exceptional research engineers that can push the boundaries of our frontier models. Specifically, we are looking for those that will help us shape our empirical grasp of the whole spectrum of AI safety concerns and will own individual threads within this endeavor end-to-end. You will own the scientific validity of our frontier preparedness capability evaluations—designing new evals grounded in real threat models (including high-consequence domains like CBRN as well as cyber and other frontier-risk areas), and maintaining existing evals so they don’t stale or silently regress. You’ll define datasets, graders, rubrics, and threshold guidance, and produce auditable artifacts (evaluation cards, capability reports, system-card inputs) that leadership can trust during high-stakes launches.
Responsibilities
Work on identifying emerging AI safety risks and new methodologies for exploring the impact of these risks
Build (and then continuously refine) evaluations of frontier AI models that assess the extent of identified risks
Design and build scalable systems and processes that can support these kinds of evaluations
Contribute to the refinement of risk management and the overall development of 'best practice' guidelines for AI safety evaluations
Qualifications
Minimum
Are passionate and knowledgeable about short-term and long-term AI safety risks
Demonstrate the ability to think outside the box and have a robust 'red-teaming mindset'
Have experience in ML research engineering, ML observability and monitoring, creating large language model-enabled applications, and/or another technical domain applicable to AI risk
Are able to operate effectively in a dynamic and extremely fast-paced research environment as well as scope and deliver projects end-to-end
Preferred
First-hand experience in red-teaming systems—be it computer systems or otherwise
A good understanding of the (nuances of) societal aspects of AI deployment
Excellent communication skills and the ability to work cross-functionally