Set and evolve Neo’s research agenda, deciding where our researchers, engineers, compute, and external collaborations can have the greatest impact.
Lead major research projects yourself, from problem selection and methodology through analysis and publication.
Develop rigorous approaches to studying difficult-to-observe frontier-model behaviours, including cases where models may recognize evaluations or have incentives to conceal relevant behaviour.
Guide and mentor researchers, review experimental designs and results, and help maintain a high bar for evidence across Neo’s work.
Work closely with research engineers to turn ambitious research questions into robust experimental systems.
Build collaborations with AI Safety Institutes, frontier labs, academics, and model developers, and represent Neo’s research externally.
Help recruit exceptional researchers and shape Neo’s research culture as the organization grows.
Qualification
You should haveHelpfulA PhD or equivalent research outputResearch on dangerous capabilities
Required
You may be an established research leader, or an exceptional researcher ready for your first opportunity to lead a serious agenda. In either case, we care most about research judgment: the ability to recognise important questions, choose tractable approaches, and produce rigorous work.
You should have:
A track record of original research in AI safety, model evaluations, machine learning, or a closely related field.
Strong understanding of frontier-model behaviour and evaluation or elicitation methodology.
The ability to turn broad problems into concrete research programmes and published results.
Excellent technical writing and the ability to communicate with researchers, engineers, and external partners.
Comfort operating in an early-stage environment where the important questions are not fully settled.
A collaborative working style and willingness to stay hands-on.
Helpful, but not required:
A PhD or equivalent research output.
Research on dangerous capabilities, alignment, agentic systems, or model evaluations.
Experience leading research projects or building a research function.
