Thesis Opportunity Spring 2026
Create AI powered quality control with AI & computer vision
Do you want to apply AI and computer vision to one of the most challenging environments in heavy industry? At Combi Wear Parts, you’ll get the chance to turn cutting-edge research into real industrial impact. This thesis project gives you the opportunity to modernize quality control at one of the Nordic region’s leading steel foundry groups.
Mission
Automated defect detection in cast components
Quality control is a critical part of the foundry process – but today it is still largely manual, time-consuming and dependent on subjective evaluation. Your mission is to develop and train an AI-based computer vision system that automatically detects surface defects on cast metal components. This thesis combines state-of-the-art technology with hands-on industrial application. You’ll start from raw image data and end with a validated prototype tested in real production environments.
You will work with:
-
Mapping and defining the most critical defect types
-
Collecting and organizing image data from production
-
Annotating images and exploring the use of synthetic datasets
-
Training and evaluating open-source computer vision models (object detection, segmentation and more)
-
Testing and validating the system using industrial cameras on-site
This is a rare chance to work with advanced AI while being deeply involved in a live production process.
Who are we looking for?
You are studying a bachelor’s or master’s degree within fields such as computer engineering, AI, machine learning, mechatronics or similar. You have a strong interest in computer vision and industrial problem-solving, and you enjoy working both analytically and practically – in the lab and directly on the factory floor.
Practical Information
Period: January – June 2026 (depending on thesis level)
Location: Flexible, with on-site practical work at the foundry in Ljungby
Level: Master recommended, but bachelor’s thesis projects are also welcome
Supervisors:
Combi Wear Parts: Anton Ekman
RISE: Andreas Thore & Lennart Elmquist
Interested?
Send your application or questions to:
