Projects

Current. I lead work on LLM value-alignment training and benchmarking, including a human-annotated benchmark of approximately 10,000 instructions across commonsense moral reasoning, deontology, justice, utilitarianism and virtue ethics, independently annotated by three annotators and prepared for open release through CLARIN.SI.

Current. Research in theology, philosophy, ethics, psychology and computer science on identity, embodiment, empathy, vulnerability, spirituality and moral agency in digital environments. The project develops ethical and theological accounts of these changes and guidance for human-centred technology.

Upcoming. EQUAL-AI proposes a European network of STEM Skills Foundries linking higher education, VET, industry, innovation infrastructure and entrepreneurship.

Current collaboration. Research on the ethical and theological dimensions of AI-mediated digital culture, including virtue, identity, human dignity and relationships. The programme examines how digital technologies shape human experience and develops ethical and theological frameworks for engaging with digital life.

Current collaboration. Research on large language models and language technologies for digital-humanities tasks, including OCR and the processing of complex Slovene cultural-heritage texts.

Current collaboration. Research programme addressing digital transformation and intelligent technologies for smart public governance.

Completed. Developed a privacy-preserving cardiac-monitoring pipeline using wearable ECG, smartphone PPG, inertial context, on-device inference and federated learning.

Completed. Developed a dynamic, resource-efficient machine-learning framework for UAV-based precision agriculture, including context-aware model adaptation, real-time weed detection and efficient deployment on resource-constrained platforms. The project later produced multiple publications and received the 2023 Agrobiznis Agro Hi-Tech 'Best idea' award.

Research on context-aware adaptation and approximation for resource-constrained mobile and edge systems, contributing to work on runtime efficiency, mobile inference and adaptive computation.

Research on resource-efficient mobile computing and approximate computation, forming part of the technical foundation for later adaptive mobile-AI work.

Developed an embodied avatar-based conversational agent with multimodal interaction and dialogue management for deployment as a virtual museum guide.

Developed a multispectral imaging platform and processing pipeline for multi-channel acquisition with applications in cultural heritage, material characterization and environmental analysis.