PhD position Reinforcement learning for digital finance

The successful applicant will join the Industrial Engineering and Business Information Systems (IEBIS) section of the High-Tech Business & Entrepreneurship Department (HBE) at the Faculty of Behavioural, Management and Social Sciences (BMS).

Background
This Ph.D. position is one four positions at the University of Twente (UT) and one of 19 positions in the context of the international Marie Sk?odowska-Curie Actions project DIGITAL. For the general description of DIGITAL and its Ph.D. positions, please check this page. Information about all other positions is available at EURAXESS, if you would be interested in any of the other positions as well, clearly state that in your cover letter.

DIGITAL' main goal? To significantly advance the methodologies and business models for Digital Finance through the use of five interconnected research objectives:

1. Ensure sufficient data quality to contribute to the EU's efforts to build a single digital market for data;
2. Address deployment issues of complex artificial intelligence models for real-world financial problems;
3. Validate the utility of pioneering eXplainable Artificial Intelligence (XAI) algorithms to financial applications and extend existing frameworks;
4. Design risk management tools concerning the applications of Blockchain technology in Finance;
5. Simulate financial markets and evaluate products with a sustainability component.

The challenge
Reinforcement Learning (RL) has become a popular paradigm for automating decision-making under uncertainty in complex environments. Although deep RL has had several breakthroughs in recent years and proven impressive algorithmic performance in closed environments, it has not yet found its way to real-world applications in open environments. In practice, RL algorithms have to work with imperfect data, be integrated into existing ecosystems, and be of use to human decision-makers. Additionally, the financial sector is subject to heavy regulation and high standards concerning risk management, fairness, and explainability. Although successful integration of RL may enhance the quality of decision-making in digital finance, several hurdles need to be overcome. Thus, this Ph.D. project examines how RL can advance digital finance.

You will address several RL implementation issues in digital finance, including both technical challenges and domain-specific ones. Utility-based RL results will improve financial decision-making by developing multi-criteria analysis, extreme scenarios, and risk management methods. RL in decision support will be optimized for explainability, regulatory compliance, model abstractions, and human judgment. We will also examine technological challenges like digital twin environments, machine learning pipelines, and digital finance ecosystem integration.

Requirements:

We look for a highly motivated, enthusiastic researcher who is driven by curiosity and has/is:

General skills:

• Master’s degree or equivalent experience in Business, Economics, Engineering, Information Technology, Computational Science or related fields;
• A strong passion and outstanding skills in data science and experience working with programming languages and statistical software such as Python, R, Stata or SAS;
• Knowledge of quantitative modeling of financial markets, econometric techniques, machine learning, or quantitative empirical research methods;
• The ability to work on real-world problems in an interdisciplinary and internationally oriented environment;
• Good communication skills and an excellent command of English.

Project-specific skills:

• Good programming skills. Experience with machine learning libraries (e.g., TensorFlow, PyTorch, Jax) is helpful;
• Strong mathematical background, in particular statistics & probability, linear algebra, calculus, optimization, and Markov decision processes;
• Knowledge of data engineering concepts such as machine learning pipelines, product deployment, and data ingestion;
• Interest in practical applications of reinforcement learning, with attention to aspects such as explainability and regulation;
• Familiarity with reinforcement learning (machine learning) is a plus.

Interested and motivated candidates are encouraged to apply, even when not yet possess all desired skills. Through self-driven learning and doctoral training, you will be able to develop relevant skills on the job.

Secondments at Industry and the European Central Bank
This Ph.D. position includes two research stays at industrial partners: Provisionally, you will be 18 months at CardoAI in Milan, Italy. Second, you will have six months at the European Central Bank (ECB) in Frankfurt, Germany. During those secondments, you will gain hands-on experience on industry-relevant research projects related to applications of AI in finance. Precise details are subject to change.

Salary Benefits:

We encourage high responsibility and independence while collaborating with colleagues, researchers, other university staff, and partners. We follow the terms of employment by the Dutch Collective Labour Agreement for Universities (CAO). Our offer contains: a full-time 4-year Ph.D. position with a qualifier in the first year; excellent mentorship in a stimulating research environment with excellent facilities; and a personal development program within the Twente Graduate School. It also includes:

• Gross monthly salary of € 2.770 in the first year, increasing each year up to € 3.539 in the fourth year;
• Excellent benefits including a holiday allowance of 8% of the gross annual salary, an end-of-year bonus of 8.3%, and a solid pension scheme;
• 29 holidays per year in case of full-time employment;
• A training program as part of the Twente Graduate School and the European MSCA Doctoral Training Programme on Digital Finance, where you and your supervisors will settle a plan for a suitable education and supervision;
• A green campus with free access to sports facilities and an international scientific community;
• A family-friendly institution that offers parental leave (both paid and unpaid);
• A full status as an employee at the UT, including pension, health care benefits, and good secondary conditions are part of our collective labor agreement CAO-NU for Dutch universities.

38 - 40 hours per week

Drienerlolaan 5