Universität Basel
Basel
Postdoc Position in Recursive Self-Improvement (Prof. I. Bogunovic)
- 18 novembre 2026
- 100%
- 4000 Basel
À propos de cette offre
The research group Rhine AI of Prof. I. Bogunovic at the Department of Mathematics andComputer Science at the University of Basel is inviting applications for a Postdoc positionin recursive self-improvement (RSI) and open-ended learning.
Today's most capable AI systems still depend on people to curate their data, build theirtraining environments, and decide what they should learn next. RSI asks whether AIsystems can increasingly close this loop themselves: generating their own tasks andenvironments, checking their own progress, and improving the very procedures used totrain them, so that each round of improvement enables the next. We see this as one of themost consequential open questions in AI, and we want to understand what makes open-ended self-improvement work, where it stalls, and how to keep it reliable and aligned.
Today's most capable AI systems still depend on people to curate their data, build theirtraining environments, and decide what they should learn next. RSI asks whether AIsystems can increasingly close this loop themselves: generating their own tasks andenvironments, checking their own progress, and improving the very procedures used totrain them, so that each round of improvement enables the next. We see this as one of themost consequential open questions in AI, and we want to understand what makes open-ended self-improvement work, where it stalls, and how to keep it reliable and aligned.
Your position
The successful candidate will lead research on how large language models (LLMs) and AIagents can drive their own improvement, and help shape the group's agenda in this area.Research directions include:
The starting date is flexible.
Your tasks include:
- Post-training LLMs with reinforcement learning for long-horizon, multi-stepreasoning and agentic tasks
- Frontier learning: an open-ended post-training approach that keeps generating newtraining problems at the edge of a model's capabilities and tracks that edge as themodel improves
- Building RL environments for agentic tasks such as search, terminal use, and games,including synthetic environments, tasks, and rewards that LLMs generate and verifythemselves
- Weight-space programming: agents that improve open-weight models by actingdirectly on their weights, opening a path to models that improve themselves
- World models that let agents plan, imagine, and learn from simulated experience
- Open-ended and evolutionary search over programs, agents, and training recipes
The starting date is flexible.
Your tasks include:
- Take a leading role in shaping and driving the group's research on recursive self-improvement and open-ended learning
- Develop new methods and run large-scale experiments with state-of-the-art openLLMs and agents
- Mentor PhD students and help coordinate projects across the group
- Write research papers and present results at leading international conferences. Wepublish in venues such as ICML, NeurIPS, and ICLR
- Interact with national and international teams of AI and machine learning experts, aswell as industrial collaborators
Your profile
Applicants must hold a PhD in machine learning or a related field, with a strongpublication record at leading machine learning venues. Candidates should have a solidmathematical foundation and strong coding and implementation skills, with hands-onexperience running large-scale experiments.
Experience with LLM post-training, reinforcement learning, world models, generativemodeling, or evolutionary and open-ended methods is a strong plus. We are looking forindependent, ambitious researchers who want to help define a new research area.
Excellent communication skills and fluency in English (spoken and written) are required.
We also seek to increase the number of women and other groups in areas where they areunderrepresented and therefore explicitly encourage such groups to apply.
Experience with LLM post-training, reinforcement learning, world models, generativemodeling, or evolutionary and open-ended methods is a strong plus. We are looking forindependent, ambitious researchers who want to help define a new research area.
Excellent communication skills and fluency in English (spoken and written) are required.
We also seek to increase the number of women and other groups in areas where they areunderrepresented and therefore explicitly encourage such groups to apply.
We offer you
The position is based in Basel, a lively city where the borders of Switzerland, France, andGermany meet. In addition, we offer:
- The chance to lead research on one of the most consequential open questions in AI,with considerable independence to pursue ambitious ideas
- Substantial compute for large-scale experiments, including the lab's own GPU serversand access to national supercomputing facilities
- Opportunities to mentor PhD students, shape the group's research agenda, and buildyour own profile toward an independent career in academia or industry
- Close collaboration with national and international collaborators
- A growing group supported by a recent ERC Starting Grant and the Swiss NationalScience Foundation (SNSF)
- The possibility of traveling to international conferences to meet distinguishedscientists
- Excellent salary, benefits, and working conditions
Application / Contact
If interested, please send us an application as described below. Your application shouldinclude two PDF files containing the following:
If interested, please send us an application as described below. Your application shouldinclude two PDF files containing the following:
- Short motivation statement (maximum 150 words) explaining your interest in theposition, which research directions you would like to lead, and why your backgroundis a good fit
- Curriculum vitae, including a list of publications
- Contact information (no direct recommendation letters) for people who canrecommend you
- Your PhD thesis or up to two representative publications
Universität Basel
4000 Basel
4000 Basel