His research focused on the use of machine learning, particularly deep learning methods, for image and video analysis. In video analysis, he developed deep learning architectures for egocentric videos that integrated both action appearance and motion within a single model, implementing noise injection as a regularization technique during training. The development and evaluation were carried out using publicly available datasets.

Additionally, he has collaborated with the Life Sciences and Veterinary Sciences departments at the university on several machine-learning projects. This interdisciplinary work involved developing and applying advanced machine learning techniques to solve complex problems in biological and veterinary sciences.