2 doctoral researchers for reinforcement learning

At Aalto University, science and art meet technology and economics. We build a sustainable future by achieving breakthroughs in our industries and their intersections, inspiring future changemakers, and creating new solutions to the world's great challenges. Our community includes 16,000 students and 5,200 employees, of whom 446 are professors. Our campus is located in Espoo. Diversity is part of our identity, and we continuously work to promote the diversity and inclusion of our community. Therefore, we warmly encourage qualified applicants from all backgrounds to join our community.

We are currently looking for two highly motivated doctoral researchers for the Cyber-Physical Systems research group at Aalto University School of Electrical Engineering for an ERC Starting Grant-funded project, which focuses on continual reinforcement learning in real-world systems. The funding offers a stable, well-resourced research environment for four years. The research focuses on real-world systems to which classical reinforcement learning theory does not apply: systems that do not return to the initial state, where errors can be irreversible, and which operate in changing environments.

Traditional reinforcement learning aims to solve the problem by optimizing the average over many hypothetical futures. In the presence of non-ergodic dynamics, such an average can deviate arbitrarily from what a single agent actually observes over a single trajectory. Typically, reinforcement learning produces a time-invariant solution that is unable to adapt to a changing environment. This project aims to solve both challenges simultaneously by combining trajectory-centric objectives with adaptive, context-dependent strategies. This enables, for example, autonomous vehicles to adapt to seasonal changes, production systems to adapt to demand without shutdowns, and medical monitoring to be personalized for patients over decades.

The selected doctoral researchers will develop fundamental theory and practical algorithms that bring reinforcement learning closer to real-world systems. The first doctoral researcher will focus on trajectory-centric optimization, the second on adaptive control principles, and both will work together to combine the approaches into a single reinforcement learning algorithm. Although the work is primarily methodological and theoretically oriented, Aalto's robotics laboratory offers the opportunity to experimentally evaluate algorithms, for example, with robotic arms or quadruped robots.

Research Focus

This project lies at the intersection of two problems: combining ergodicity theory with reinforcement learning is a fresh research direction, and how reinforcement learning agents can safely observe and adapt to a changing environment is a timely and active field of research. Combining trajectory-centric objectives with adaptive, context-aware control principles is a new research area. The doctoral researchers will advance it from these two complementary perspectives and work closely together to combine both into one algorithm:

development of trajectory-centric stochastic optimization theory under non-ergodic dynamics;

implementation of practical reinforcement learning algorithms that optimize the long-term performance of individual agents;

development of change detection algorithms to infer when a solution is no longer valid and needs to be adapted;

implementation of efficient, provably safe solutions;

combining these advancements into state-of-the-art reinforcement learning algorithms;

validation of algorithms in relevant simulation environments and hardware experiments.

Requirements

Qualified applicants are required to have:

a higher university degree in computer science, mathematics, electrical engineering, or a related field;

fluent oral and written English language skills;

the ability to work both independently and as part of a research group.

If you are selected for the position, you will apply for study rights in the Aalto University School of Electrical Engineering doctoral programme. See student information and selection criteria at https://www.aalto.fi/en/study-options/aalto-doctoral-programme-in-electrical-engineering .

Desired Background

We are looking for applicants with a strong background in one or more of the following areas:

stochastic processes, preferably also ergodicity theory;

reinforcement learning, dynamic optimization, and Markov decision processes;

programming skills (Python).

We Offer

A fully funded doctoral researcher position at Aalto University, which is consistently ranked among Europe's best universities.

A fixed-term position following the university's standard 2+2 model. The employment is initially for two years, including a six-month probationary period. The employment will be extended by two additional years after a successful mid-term evaluation, making the total duration four years.

The work begins in January 2027 or as agreed.

The doctoral researcher's starting salary is 3143 €/month.

The opportunity to work in a research project where your work has an impact.

You will become part of a new and dynamic research group with plenty of opportunities for collaboration and exchange of ideas. The group has strong international connections and active research collaboration with, among others, RWTH Aachen University, KTH (Stockholm), and the London Mathematical Laboratory. During doctoral studies, it is also possible to make research visits to these and other partners.

Interested?

You can apply for the position through our recruitment system (“Apply now!” at the bottom of the page) by 23.10.2026 at 23.59 (EEST). Please attach the following materials in English and in PDF format:

A motivation letter, in which you also mention which of the two positions you are applying for;

A curriculum vitae (CV), which includes contact information for at least two referees;

Copies of bachelor's and master's degree certificates and transcripts of records.

Please note: Current employees of Aalto University must apply for the position through the Workday system (Internal Jobs) using their own Workday user ID (not through the external open jobs page). If you are an Aalto University student or visitor, apply with your personal email address (not an aalto.fi address) through the Aalto University open jobs page.

For more information about the positions, contact Dominik Baumann ( dominik.baumann@aalto.fi ). For questions regarding the application process, contact HR advisor Johanna Haapalainen ( hr-elec@aalto.fi ).

We review applications and may invite suitable applicants for an interview during the application period. We strive for a transparent and equitable recruitment process, so please feel free to ask us for feedback.

Do you want to know more about us and your future colleagues? You can watch these videos:

This is Aalto University!

Aalto University – Towards a better world

and Shaping a Sustainable Future .

Read more about working at Aalto: https://www.aalto.fi/en/careers-at-aalto and explore our new virtual campus experience: https://virtualtour.aalto.fi