Research
My research focuses on machine learning for healthcare, especially work that is interpretable, practical, and easy to connect to real-world use. My strongest area is biomedical time-series learning, and I also work on medical imaging, multimodal learning, and related applied AI problems.
Current research (Research Assistant, Malmö University)
At Malmö University, I work as a Research Assistant on an EU-funded research project in interpretable biomedical time-series learning. The main problem I work on is epileptic seizure prediction using signals such as ECG and EEG.
My current work looks at how we can build models that are not only accurate, but also easier to understand. In simple terms, I am interested in models that can show why a decision was made, instead of acting like a black box. This is especially important in healthcare, where doctors and researchers need systems they can trust.
I mainly work with prototype-based deep learning, attention mechanisms, and representation learning. The goal is to build systems that can support better prediction while also providing explanations that are meaningful in clinical settings.
CREME VIP research programme
In the CREME VIP research programme, I worked on radar-based people counting for noisy real-world environments. The core challenge in this work was that radar data can be affected by clutter, noise, and changing surroundings, which makes counting people more difficult in practice.
This project combined signal preprocessing, feature extraction, and machine learning to improve counting accuracy. A major strength of this work was keeping the system focused on practical deployment, not only on laboratory performance.
Research Assistant, Green Networking Research Group, CSEDU
I also worked as a Research Assistant in the Green Networking Research Group, CSEDU, where I was involved in research on energy-aware wireless sensor networks. This work focused on how to improve network lifetime and anchor selection in systems that depend on limited and harvested energy.
That research led to a publication in IEEE R10-HTC. It gave me a strong foundation in research thinking, problem formulation, experiments, and technical writing. It also helped shape my long-term interest in building efficient and trustworthy systems.
Research interests
My broader research interests include machine learning, biomedical AI, ECG/EEG time-series learning, medical imaging, multimodal learning, LLMs, generative AI, and interpretable ML. I am especially interested in work that connects strong machine learning methods with useful real-world impact.
Publications
Network Lifetime Aware Anchor Selection for Energy Harvesting Wireless Sensor Networks