We’re seeking multiple research fellows who can work with our other research fellows to solve open problems related to superintelligence alignment, and prepare those results for publication. For those with some graduate study or a Ph.D. in a relevant field, the salary starts at $65,000 to $75,000 per year, depending on experience. For more senior researchers, the salary may be substantially higher, depending on experience. All full-time employees are covered by our company health insurance plan. Visa assistance is available if needed.
Ours is a young field. Our current research agenda includes work on tiling agents, logical uncertainty, decision theory, corrigibility, and value learning, but those subtopics do not exhaust the field. Other research topics will be seriously considered if you can make the case for their tractability and their relevance to the design of self-modifying systems which stably pursue humane values.
This is not a limited-term position. The ideal candidate has a career interest in these research questions and aims to develop into a senior research fellow at MIRI, or aims to continue these avenues of research at another institution after completing substantial work at MIRI.
Some properties you should have
- Published research in computer science, logic, or mathematics.
- Enough background in the relevant subjects (computer science, logic, etc.) to understand MIRI’s technical publications.
- A proactive research attitude, and an ability to generate productive new research ideas.
A formal degree in mathematics or computer science is not required, but is recommended.
Our recommended path for becoming a MIRI research fellow
Unfortunately, no university yet offers advanced degrees in our field of superintelligence alignment, and few people that we don’t already know publish work in the field. Thus, we are unlikely to hire researchers on the basis of their past academic history alone.
Therefore, to demonstrate the research ability we seek in research fellow hiring candidates, applicants should aim to do one or more of the following things:
- Attend one or more MIRI research workshops. Apply here.
- Attend several meetings of the MIRIx group nearest you.
- Study our research thoroughly enough that you can publish interesting new technical contributions on our IAFF forum.
MIRI’s mission is to ensure that the creation of smarter-than-human intelligence has a positive impact. Our strategic focus is our technical research agenda on “superintelligence alignment,” composed of numerous subproblems in AI, logic, decision theory, and other fields.
Our technical research program currently employs four full-time research fellows, fosters collaboration with our research associates and others, runs several research workshops each year, and funds independently-organized MIRIx workshops around the world.
Our offices are located in downtown Berkeley, two blocks from the UC Berkeley campus. We work flexible hours, and sometimes work from home, though working from the office is nice because it has free meals, snacks, and drinks.
Before applying for a fellowship, you’ll need to have attended at least one research workshop.