Dr. Alexander Gutev is a researcher and technology professional specializing in Artificial Intelligence, Machine Learning, Computer Vision, and Human Action Recognition. His doctoral research at the University of Malta focused on computationally efficient methods for human action recognition using 3D video content, with particular emphasis on motion-saliency detection, multimodal fusion, and privacy-preserving depth-based representations. His research has contributed to improving the efficiency of video-based learning models by selectively identifying informative motion regions, reducing training and inference requirements.

He holds a PhD in Communications and Computer Engineering from the University of Malta (2021–2025), along with an MSc in Computer Science and a BSc (Honours) in Computing Science. His technical expertise includes deep learning, anomaly detection, multimodal fusion, time-series analysis, image and video processing, object recognition, segmentation, and real-time analytics. He is proficient in Python, C/C++, PyTorch, OpenCV, NumPy, scikit-learn, CNNs, RNNs, Transformers, CUDA, and other machine learning technologies.

Dr. Gutev has authored several peer-reviewed IEEE publications on human action recognition, motion saliency, depth-based analysis, and object tracking. He has presented his research at international conferences including IEEE EUVIP, AICCONF, ICCP, MELECON, ELMAR, and EUROCON. He has also contributed to the academic community through peer-review activities for conferences and journals including IEEE VCIP, IEEE ICME, Information Sciences, and Engineering Applications of Artificial Intelligence.

Alongside his research career, Dr. Gutev has strong industry experience. He currently serves as Project Manager at Viditrack Ltd, where he manages software development projects and multidisciplinary teams. Previously, he worked as a Lead Developer at CG Algorithms Ltd, contributing to software, mobile applications, backend systems, and algorithm development.

His combination of AI research, computer vision expertise, software engineering, project leadership, and international publication experience makes him a strong contributor to the development and practical application of intelligent systems and emerging AI technologies.

Vision-Centric AI for Human-Aware Intelligent Systems:

From Activity Recognition to Human-AI Collaboration Artificial intelligence systems that can perceive and understand human behaviour are becoming increasingly important for healthcare, assisted living, robotics, and human-AI collaboration. Computer vision provides a key foundation for these systems by enabling machines to interpret visual and spatial information about people and their interactions with the surrounding environment. However, developing reliable vision-based AI for real-world applications requires not only accurate recognition, but also computational efficiency, privacy awareness, and the ability to operate effectively in complex environments. This presentation explores research into computer vision and machine learning for human activity recognition using 3D video data, with a focus on developing intelligent systems capable of understanding human behaviour. The work investigates motion saliency techniques that identify the most informative regions of video and selectively focus computational resources on relevant information. These approaches are combined with depth, colour, and skeleton-based representations to develop robust multimodal activity recognition systems. Beyond activity classification, this research provides a foundation for AI systems that can perceive and respond to human behaviour in a meaningful manner. Potential applications include healthcare and assisted living, where vision-based systems could support unobtrusive monitoring of activities, assist with the recognition of unusual events, and provide contextual information to human caregivers. The use of depth and skeleton information also provides opportunities for developing more privacy-aware alternatives to conventional RGB-based monitoring. The presentation will further discuss how efficient computer vision can contribute to human-AI collaboration by allowing intelligent systems to extract meaningful information from complex environments while reducing unnecessary computational processing. These principles are relevant to emerging intelligent systems, including autonomous and assistive technologies, in which AI must operate alongside humans rather than simply perform isolated recognition tasks. Overall, the work demonstrates how advances in computer vision and multimodal machine learning can contribute to the development of more capable, efficient, and human-aware AI systems, with applications extending from healthcare and assisted living to broader forms of intelligent automation and human-AI collaboration