About the role
You turn the ambition of personal intelligence into questions that can be answered honestly, and connect foundational research to capabilities a person can understand and control. You will lead work on efficient models, continual learning, planning, privacy and human collaboration, and establish the reproducible baselines and independent evaluation that keep the rest of us truthful.
The work
Lead research in efficient models, continual learning, planning, privacy and human collaboration. Establish reproducible baselines, independent evaluation and a publication process that respects user data and collaborators' rights. Work directly with product and systems leaders on which hypotheses deserve compute.
What good looks like
In your first 90 days, establish three bounded research programs, reproduce their baselines and propose explicit continuation or stop criteria.
Evidence we look for
Bring a sustained record of original research, rigorous experiments and mentorship. A doctorate or equivalent research contribution is relevant; the strength of your reasoning matters more than an institution's name.
What we need to see
- A sustained record of original research, with rigorous experiments behind it
- You have mentored researchers, and can name what they went on to do
- You set up evaluation that is honest about negative results rather than optimised to look good
- You can choose which questions are worth years, and defend the choice
Nice to have
- A doctorate, though the strength of your reasoning matters more than the institution
- Efficient models, continual learning, or privacy-preserving machine learning specifically
- You have run a research group as well as done research
The exercise
Present one important result, its strongest alternative explanation and an experiment that could disprove your current thesis.
Where and how we work
In the office together five days a week, in any of these cities. Remote-friendly around your family, arranged one person at a time.