Research Fellow (Mathematics)

Job Description

The successful candidate will work with Asst. Prof. Julian Sester on developing mathematical methods for Distributionally Robust Reinforcement Learning.

The main responsibilities of the position include:

• Developing rigorous mathematical foundations for distributionally robust reinforcement learning under model uncertainty;
• Studying robust Markov decision processes, dynamic programming principles, and convergence properties of robust learning algorithms;
• Designing and analysing reinforcement learning algorithms based on Wasserstein, Sinkhorn, or related ambiguity sets;
• Establishing theoretical guarantees such as stability, convergence, approximation error bounds, or sensitivity estimates;
• Implementing and testing proposed methods in numerical case studies, with possible applications in quantitative finance, stochastic control, or risk management;
• Preparing research manuscripts for publication in leading journals and presenting results at seminars, workshops, and conferences;
• Contributing to the research environment of the group, including discussions with PhD students, postdoctoral researchers, and collaborators.

Qualifications / Discipline:

• PhD in Mathematics, Applied Mathematics, Statistics, Operations Research, Quantitative Finance, or a closely related discipline.

Skills:

• Strong background in probability theory, stochastic processes, stochastic control, optimization, or reinforcement learning;
• Solid mathematical training and ability to work with rigorous proofs;
• Familiarity with Markov decision processes, dynamic programming, distributionally robust optimization, optimal transport, or reinforcement learning is highly desirable;
• Programming skills in Python are desirable, especially experience with numerical experiments, machine learning libraries, or reinforcement learning environments;
• Good written and oral communication skills;
• Ability to work independently and collaboratively in an interdisciplinary research environment.

Experience:

• Prior research experience in one or more of the following areas is desirable: reinforcement learning, robust control, stochastic control, distributionally robust optimization, optimal transport, mathematical finance, or machine learning;
• A track record of high-quality research, demonstrated through publications, preprints, or a strong PhD thesis;
• Experience with numerical implementation of mathematical or machine learning methods would be an advantage;
• Experience in quantitative finance or financial applications is welcome but not required.