Research Scientist

Resolution•Berkeley, CA

5d•Onsite

About The Position

Resolution conducts research focused on aligning artificial superintelligence (ASI). The primary goal is to ensure that ASI, which may emerge in the near future, is aligned with human values. The organization aims to achieve a higher a priori confidence in aligned outcomes by exploring a diverse range of theoretical and empirical approaches. Significant investment is made in research automation to accelerate progress, with the belief that stronger alignment theory leads to more effective automation. Resolution was established in 2026 by former members of the UK AISI's Alignment Team and Timaeus, pioneers in applying singular learning theory to alignment.

Requirements

A graduate degree (Ph.D.) or equivalent experience in a field relevant to the program (ML, CS, mathematics, physics, statistics, philosophy, or closely related).
A track record of research and strong technical writing ability (papers, preprints, or comparable output).
A strong mathematical background, even if work is primarily empirical.
Ability to credibly articulate why their area of alignment work is worth pursuing.
Willingness to use AI tools aggressively in their workflow with appropriate caution.
Motivation by the alignment of artificial superintelligence (ASI) and a desire to contribute full-time.

Nice To Haves

For scalable oversight: Hands-on experience with debate, prover-verifier games, RLHF empirics at frontier scale; familiarity with the scalable oversight literature; background in game theory, complexity theory, mechanism design, or decision theory.
For complexity theory: Demonstrated expertise in theoretical computer science, including creative modeling and assumption generation.
For learning theory: Background in algebraic geometry, Bayesian statistics, information theory, statistical physics, optimization theory, or learning theory; familiarity with (singular) learning theory.
Across programs: Experience scaling experiments to billion-plus parameter models; a track record of productive collaboration with engineers; prior engagement with Resolution's research or sibling organizations (Simplex, ARC, etc.).
Experience with AI alignment is a plus, but not a requirement.

Responsibilities

Conduct research within a designated program, translating and modeling alignment problems into concrete empirical or theoretical forms and executing on these problems.
Design and run experimental protocols and methods for empirical research, or prove, conjecture, and invest in autoformalization for theoretical research.
Write and present completed research in the form of papers, blog posts, and talks.
Communicate research progress and obstacles to team members daily via Slack and in weekly meetings.