AIXI Labs Research Fellowship

AIXI Labs is a small, London-based non-profit doing AI safety research grounded in algorithmic information theory, fiscally sponsored by the 501(c)(3) nonprofit Principles of Intelligence. Most AI research today is either mathematically clean but narrow, or practical but too opaque for rigorous safety analysis. We target a third category: methods general enough to describe powerful agents and analyzable enough to support real safety claims — using AIXI, the theoretical model of unbounded artificial superintelligence — then port the strongest of these ideas to modern LLM-based agents.

Our fellowship is a quarterly, rolling program for researchers who want to work on this agenda under the close mentorship of our core team — Cole Wyeth (Founder & Executive Director), Marcus Hutter (Research Director, part-time), and Aram Ebtekar (Founding Research Scientist). Fellows pursue a focused research project over roughly a quarter, and receive a monthly stipend plus a compute budget for machine-learning work. The fellowship is fully remote — you may work from anywhere in the world, or from our London base. Applications are reviewed on a rolling basis, and previous applicants remain eligible for later rounds. In exceptional cases, fellowships may be renewed.

Minimum qualifications

• A research background at the level of a current PhD student. For primarily empirical projects, a master’s-level command of the relevant mathematics paired with substantial industry or applied research experience can also qualify.
• Solid grounding in probability, statistics, information theory, and reinforcement learning, sufficient to engage seriously with AIXI-style theoretical models. Prior expertise in algorithmic information theory or AIXI is not required.
• Strong verbal and written communication skills.
• Genuine motivation to reduce existential risk from advanced AI, and enthusiasm for the kind of work we do.

Preferred qualifications

• Publications at top ML/AI/theory venues (e.g. NeurIPS, ICML, ICLR, ALT, COLT), though we value the strength of your best work over volume.
• Background in algorithmic information theory, computability theory, learning theory, or related theoretical computer science.
• Experience fine-tuning, evaluating, or red-teaming LLM-based and/or RL agents.
• Prior engagement with the UAI/AIXI community or related fields — including members of our online reading groups.

Responsibilities

• Pursue a self-directed research project aligned with our algorithmic-information-theoretic approach to AI safety, spanning anything from foundational theory to empirical tests with LLM-based agents.
• Shape the direction of your project around your own strengths and interests — we support a range of work within this framework.
• Contribute papers, blog posts, or other written output that communicates your results to the broader AI safety and ML research communities.