etjob.dk
Vejle kommune Ph.d., naturvidenskab og teknik

PhD Positions in XAI and Sustainable AI at SDU Centre for Software Technology

The Centre for Software Technology (CST), part of Maersk McKinney Moller Institute in the Faculty of Engineering at the University of Southern Denmark (SDU), invites applications for one or more 3-year PhD positions. The positions are open from January 1, 2026, and the specific start date will be agreed with the successful candidates.

Application Due Date: November 23, 2025, at 11:59 PM / 23:59 (CET/CEST)

What we offer:

The Centre values teamwork, professional diligence, enthusiasm for technology and the drive to adopt new skills and extended responsibilities. We offer an open, international team with flexible work organization and support of individual development. The group is involved in a variety of national and European projects and features a strong network of academic and industrial partners. We solve challenging research problems from real applications and implement novel software and solutions together with end users.

We are looking for candidates for two PhD projects:

1. PhD Position in XAI with Commonsense Knowledge for Robotics and Computer Vision

2. PhD Position in Sustainable AI for Enhancing Health Informatics

(Please scroll down to read more about the project descriptions for the respective PhD positions.)

What we expect:

The applicant should have a Master’s degree (MS / MSc / MTech / ME etc.) in Software Engineering, Information Technology, Computer Science, Artificial Intelligence, Data Science, Robotics or Health Informatics or any other relevant field, at the time of application into this PhD program.

Please scroll down to read more about the expectations, in terms of PhD candidate responsibilities, qualifications, required skills (mandatory), and additional skills (preferred), with respect to each PhD position.

Workplace description:

The Centre for Software Technology (CST) at SDU is a part of the new campus, SDU Vejle, home of excellence within information technology. The academic focus of the SDU Centre for Software Technology is to conduct research in interactive information technology and software engineering. The areas will be complementing each other to make way for the engineering of the next generation of reliable, intelligent and interactive software solutions. This includes understanding how AI technologies and data-informed software development with a human-centred perspective can change the engineering of products and software infrastructures.

Further Information:Please visit our website for more details; and feel free to reach out to the contact persons by email in case you have any more questions.

Project 1. - PhD Position in XAI with Commonsense Knowledge for Robotics and Computer Vision

There is growing trend towards explainable AI (XAI) today. Opaque-box models with deep learning (DL) offer high accuracy but are not explainable due to which there can be problems in interpretation, trust, error tracing etc. XAI methods offer clear-box models, often comparable to DL in accuracy, yet better in 1st-time scenarios (where the AI has not exactly encountered the given situation with pre-training earlier), as well as in counterbalancing bias & overfitting. In addition to classical XAI models, e.g. decision trees, there is the paradigm of commonsense knowledge (CSK), i.e. everyday knowledge on concepts, properties, and relationships. It is easy for humans and often the hardest for AI systems (vs. encyclopedic knowledge). In this project, we aim to investigate commonsense knowledge and other XAI paradigms, mainly in robotics and computer vision, encompassing avenues such as smart manufacturing, domestic robotics, and multipurpose robots. It thrives on earlier successful work, e.g. CSK in human-robot collaboration for vehicle assembly, and CSK for image detection with task-relevant classification. In general, this research project aims to study the benefits of XAI with CSK in robotics and computer vision, heading towards creating better next-generation AI-based robotic systems, especially useful in multipurpose robots. It can involve harnessing DL with XAI to obtain the best of both worlds in targeted applications.

PhD candidate responsibilities:

The PhD candidate will be responsible for conducting research and implementation in areas such as:

  • Infusing computer vision models (based solely on DL) with XAI, to incorporate subtle human judgement, thus aiming to bring robotic systems closer to the thresholds of human cognition
  • Designing XAI-based methods in robotics and computer vision to reduce algorithmic complexity by orders of magnitude, e.g. by tracing paths of trees and extraction from knowledge bases (KBs), as opposed to pure DL
  • Defining specific CSK-premises (in addition to CSK from existing KBs) in targeted avenues such as multipurpose robots, along with mathematical modeling and algorithmic insights
  • Counterbalancing issues such as bias, overfitting, and inexplicable errors in excessive pre-training: by making robots explicitly aware of the reasons behind their actions using XAI
  • Exploring trust in conjunction with XAI in human-robot interaction and collaboration, since robots and humans can obviously trust each other better if tasks are more interpretable
  • Considering subjective issues on the needs of humans working with robots from a CSK angle, e.g. boredom, encumbrance, comfort, and pleasure

Qualifications for application:

The candidate should have a Master’s degree (MS / MSc / MTech / ME etc.) in Computer Science or Information Technology or Software Engineering or Artificial Intelligence or Data Science or Robotics or any other relevant field, at the time of application into this PhD program.

Required skills (mandatory):

  • Well-versed with a general background of Robotics and Computer Vision
  • Excellent programmer in Java / C / Python or equivalent
  • Excellent at using Machine Learning software, e.g. PyTorch / TensorFlow / Scikit Learn
  • Highly knowledgeable in mathematical and statistical concepts
  • Proficient in English for technical writing, oral presentations, and general communication

Additional experience (preferred):

  • Very good knowledge of XAI techniques
  • Thorough understanding of KB development
  • Working with specific Robotics applications in multiple domains
  • Proficiency with various state-of-the-art Computer Vision models

Project 2. - PhD Position in Sustainable AI for Enhancing Health Informatics

In recent years, the need for more automated, contactless, and minimally invasive AI-based diagnostics in health informatics is receiving much attention. This can help to bridge doctor-patient gaps, allow doctors to focus on more challenging tasks (vs. routine procedures), and offer greater convenience to patients. It is important to gear towards sustainable health informatics to achieve fundamental computational goals of minimizing complexity, provide more cost-effective and accessible solutions, and hence encourage their widespread use. In line with these initiatives, and motivated by prior success stories, this project heads towards promoting sustainable AI to enhance health informatics. A useful paradigm in this project is that of Knowledge-Guided Machine Learning which entails infusing domain knowledge into ML methods to add specific context aiming to enhance performance. This project aims to explore the adaptation of KGML in health informatics, contributing to XAI because explainability is vital in health and medicine. Moreover, it leverages preferring simpler theories over complex ones if both give comparable levels of accuracy. Furthermore, it leverages the power of transfer learning where a hypothesis learned on a smaller dataset can be transferred to larger data. It also considers harnessing counterfactuals in AI for adding a solid theoretical foundation to empirical analysis. It poses challenges, particularly in real-world implementation, testing with real data, and widespread acceptance of novel technologies. Expected outcomes include the development of mobile apps and websites to assist various users, including healthcare professionals.

PhD candidate responsibilities:

The PhD candidate will be responsible for conducting research and implementation in areas such as:

  • Exploring a KGML in health informatics; this can entail working alongside medical experts to extract relevant knowledge, and programming using XAI methods to infuse that knowledge with ML methods, thus aiming to optimize performance and interpretability, analogous to RAG (Retrieval-Augmented Generation) in LLMs
  • Investigating multiple models for analysis, focusing on the Occam’s Razor principle of preferring simpler theories to complex ones in suitable scenarios; harnessing Transfer Learning with challenges such as ascertaining the minimum number of data samples to get maximum accuracy for adequate diagnosis
  • Studying XAI methods, e.g. counterfactuals in reasoning and knowledge graphs (KGs) based on domain expertise, to strengthen inferences drawn from data, and to reduce complexity of learning – by factual reasoning (vs. excessive pre-training), thus aiming to improve interpretability and trust to better assist medical professionals
  • Developing mobile apps and websites for contactless diagnostics and other assistance in healthcare, catering to various stakeholders, ranging from novice users to expert healthcare professionals.

Qualifications for application:

The candidate should have a Master’s degree (MS / MSc / MTech / ME etc.) in Computer Science or Information Technology or Software Engineering or Artificial Intelligence or Data Science or Health Informatics or any other relevant field, at the time of application into this PhD program.

Required skills (mandatory): -

  • Thoroughly well-versed with Health Informatics or related areas
  • Excellent programmer in Java / C / Python or equivalent
  • Excellent at using Machine Learning software, e.g. PyTorch / TensorFlow / Scikit Learn
  • Highly knowledgeable in mathematical and statistical concepts
  • Proficient in English for technical writing, oral presentations, and general communication

Additional experience (preferred):

  • Familiarity with the KGML paradigm and its implementation
  • Working with domain experts in Healthcare and Medicine
  • Very good knowledge of XAI techniques
  • Successful development of mobile apps and websites with real-world deployment

Contact persons (for both the PhD positions):

Dr. Aparna Varde, Full Professor, Centre for Software Technology, SDU Vejle (E-mail: johnterrellraymond13apva@mmmi.sdu.dkjones.dkthomas.dk)

Dr. Mikkel Baun Kjaergaard, Full Professor and Head of Education, SDU Vejle (E-mail: mbkjaustin37katieodom@greene-simmons.dkrichard-cole.dkmmmi.sdu.dkcoleman-hickman.dk)

For information about the application procedure, see the full posting on our website.

The application deadline is November 23, 2025, at 11.59 PM / 23.59 (CET/CEST)

Appointment as a PhD fellow is a 3-year salaried position, and the monthly gross salary incl. pension is 36.138 DKK. If you have relevant postgraduate experience, you may be placed on a higher salary step.

The positions are available from January 2026, at the start date will be agreed with the successful candidates.

Further information for international applicants about entering and working in Denmark. You may also visit WorkinDenmark for additional information.

Further information about The Faculty of Engineering.