AIXI Labs is inviting applications for its AI Safety Research Fellowship, offering researchers worldwide the opportunity to conduct advanced research in algorithmic information theory, artificial intelligence safety, reinforcement learning, and intelligent agent systems.
The fellowship is designed for researchers interested in developing rigorous foundations for understanding and improving the safety of advanced artificial intelligence systems.
Successful fellows will work on self-directed research projects under the mentorship of leading AI researchers while receiving a monthly stipend of between £2,000 and £3,500 and access to a dedicated compute budget for machine learning research.
The programme is available remotely worldwide or from AIXI Labs’ base in London, United Kingdom, with applications reviewed on a rolling basis.
About AIXI Labs
AIXI Labs is a London-based non-profit research organisation focused on advancing AI safety through the foundations of algorithmic information theory.
The organisation works on addressing a major challenge in artificial intelligence research: developing methods that are both mathematically rigorous and practical enough to support meaningful safety analysis.
While much of current AI research focuses either on highly theoretical models or practical systems that are difficult to analyse, AIXI Labs aims to bridge these approaches by studying general models of intelligence and translating useful insights into modern AI systems.
The organisation’s research agenda is inspired by AIXI, a theoretical model of unbounded artificial superintelligence, and explores how these concepts can inform the development of safer and more reliable AI agents.
The AIXI Labs Research Fellowship is a quarterly, rolling programme supporting researchers who want to contribute to AI safety research through theoretical and empirical approaches.
Fellows typically undertake a focused research project over approximately one quarter while receiving close mentorship from the AIXI Labs research team.
The fellowship encourages research across a broad range of topics, including:
- Theoretical foundations of intelligent agents
- Algorithmic information theory
- Reinforcement learning
- AI safety methodologies
- Large language model (LLM)-based agent evaluation
- Empirical testing of advanced AI systems
Researchers are encouraged to shape their projects around their own expertise and interests while contributing to AIXI Labs’ broader mission.
Fellowship Benefits
Selected fellows will receive:
- Monthly stipend of £2,000–£3,500
- Research compute budget for machine learning projects
- Close mentorship from AI safety researchers
- Opportunity to publish papers and technical blogs
- Participation in a global AI safety research network
The fellowship provides researchers with the resources and flexibility needed to explore important questions surrounding advanced artificial intelligence.
Mentorship Team
Fellows will work closely with members of the AIXI Labs core research team, including:
- Cole Wyeth – Founder and Executive Director
- Marcus Hutter – Research Director
- Aram Ebtekar – Founding Research Scientist
Through this mentorship structure, researchers receive guidance on developing rigorous research projects and communicating findings to the AI safety and machine learning communities.
Research Responsibilities
Selected fellows will be expected to:
- Develop and conduct an independent research project aligned with AIXI Labs’ AI safety approach
- Explore theoretical or empirical questions related to intelligent agents
- Conduct experiments involving LLM-based agents or reinforcement learning systems where appropriate
- Produce research outputs such as:
- Academic papers
- Technical reports
- Blog articles
- Research communications
Fellows will have flexibility to determine the direction of their research while contributing to broader discussions on AI safety.
Minimum Qualifications
Applicants should have a research background equivalent to a current PhD student.
For applicants focused primarily on empirical research, a Master’s-level understanding of relevant mathematics combined with significant industry or applied research experience may also be considered.
Candidates should demonstrate:
- Strong understanding of probability
- Knowledge of statistics
- Familiarity with information theory
- Understanding of reinforcement learning
- Ability to engage with theoretical AI models
- Strong written and verbal communication skills
- Genuine interest in reducing existential risks from advanced AI
Previous experience with AIXI or algorithmic information theory is not required, although interest in these areas is valuable.
Preferred Qualifications
Applicants with the following experience are encouraged to apply:
- Publications at leading AI and machine learning venues, including:
- NeurIPS
- ICML
- ICLR
- ALT
- COLT
- Background in:
- Algorithmic information theory
- Computability theory
- Learning theory
- Theoretical computer science
- Experience with:
- Fine-tuning large language models
- Evaluating AI agents
- Red-teaming LLM systems
- Reinforcement learning agents
- Engagement with UAI/AIXI research communities
However, applicants are evaluated based on the quality and potential of their research contributions rather than publication quantity alone.
Who Should Apply?
The fellowship is suitable for:
- PhD researchers in artificial intelligence
- Machine learning scientists
- Computer science researchers
- AI safety researchers
- Reinforcement learning specialists
- Theoretical computer scientists
- Researchers exploring trustworthy and reliable AI systems
Candidates interested in developing safer approaches to advanced AI are encouraged to apply.
Application Process
Interested applicants should submit their applications through the AIXI Labs application portal.
Applications should demonstrate:
- Research background and expertise
- Relevant technical skills
- Motivation for AI safety research
- Proposed areas of interest
- Ability to communicate complex research ideas
Applications are reviewed on a rolling basis.
Previous applicants remain eligible for future fellowship rounds, and exceptional fellows may have opportunities for renewal.
Fellowship Location
The fellowship offers flexible participation options:
- Fully remote from anywhere in the world
- London-based research engagement
This global structure enables researchers from diverse locations to contribute to cutting-edge AI safety research.
Why This Fellowship Matters
As artificial intelligence systems become increasingly powerful, understanding their behaviour, limitations, and safety risks has become one of the most important research challenges of the modern era.
The AIXI Labs Research Fellowship provides researchers with a unique opportunity to explore foundational questions about intelligence while contributing to efforts aimed at making advanced AI systems safer.
With financial support, expert mentorship, and opportunities for research publication, the fellowship represents an excellent pathway for researchers working at the intersection of mathematics, machine learning, and AI safety.
Applications are currently open and accepted on a rolling basis.
