About the Thesis
The energy system in Europe is undergoing a significant transformation with an increasing share of renewable energy, creating new challenges in dynamics and balance. District heating, which annually delivers 50 TWh of energy in Sweden, plays a central role in this system. There is large potential to optimize production to support system balance, but this requires fast and smart algorithms.
The task consists in applying reinforcement learning to optimize production planning in district heating systems. You will work in the context of Sigholm's product, Aurora by Sigholm (AbS), which already optimizes production for some of the largest district heating companies in Northen Europe. Learn more about our product at https://www.aurorabysigholm.com/en.
At Sigholm, you get the opportunity to work on projects that have a real impact on the energy transformation.
Scope: 30 credits
Location: Thesis workers are welcome to sit at one of our offices in Stockholm, Gothenburg, or Västerås. There is some possibility for remote work.
Start: Spring semester 2026, exact start date is flexible.
Application deadline: To start in spring 2026, we would like your application before December 1, 2025.
About Us
At our company we have engineers, data scientists, analysts, management consultants, project managers, energy strategists & communicators - specialists in solving the major societal challenges. Most people working at Sigholm are consultants with assignments for clients in the energy industry, but we also have a team of about 20 people who work internally with the development and maintenance of our product AbS.
We have a flat organization with decentralized leadership and a high degree of transparency. We focus greatly on self-leadership, and everyone working at Sigholm has organizational responsibilities beyond their main role to contribute to a better workplace. Learn more about how we work at https://www.sigholm.se/karriar (in Swedish).
About You
We are looking for someone who wants to contribute to the energy transition through smart technology in your thesis work.
To succeed with this task, we believe that you:
- Are studying at a master's level in computer science, mathematics, or similar.
- Have taken courses in machine learning, especially reinforcement learning, Markov processes, and Deep Q-Networks.
- Are accustomed to working in Python and with some machine learning frameworks such as TensorFlow, PyTorch, Gymnasium, Ray, or Stable Baseline3.
- Work systematically and step-by-step, driven by curiosity and passion for learning.
- Are fluent in Swedish and/or English.
- Hands-on experience with reinforcement learning is highly meritorious.
We encourage doing thesis work in pairs, so apply with a partner or let us know if you are open to being matched with another applicant.
