Makes learning feel rewarding and fun.
Passionate about student development.
Always clear, concise, and insightful.
Encourages students to think independently.
Dr Andreas Shepley serves as a Lecturer in Computational Science at the University of New England (UNE), School of Science and Technology, Faculty of Science, Agriculture, Business and Law. He earned his PhD from UNE, focusing on challenges in adopting computer vision technologies for applied sciences, where he developed a novel algorithm to improve object detection accuracy and recall. Shepley published strategies for training neural networks to enhance robustness and domain invariance in ecological image processing and created U-Infuse, an open-source software tool democratizing customizable deep learning for object detection in camera trap imagery. His academic background includes a BSc (Hons) and a Graduate Degree in Information Science from the University of New South Wales (UNSW), as well as a Graduate Diploma in Education from UNE. Shepley specializes in deep learning systems, artificial intelligence, computer vision, and data science.
Currently, Shepley researches biometrics for individual wildlife recognition in collaboration with the New South Wales Department of Primary Industries. Key publications include 'Automated location invariant animal detection in camera trap images using publicly available data sources' (Ecology and Evolution, 2021), 'U-Infuse: Democratization of Customizable Deep Learning for Object Detection in Camera Trap Imagery' (bioRxiv/PMC, 2021), and 'Learning the Grid: Transformer Architectures for Electricity Price Forecasting in the Australian National Market' (Applied Sciences, 2025). He teaches courses such as Object Oriented Programming (COSC120), Deep Learning (COSC351/COSC551), data science, and machine learning. Beyond academia, Shepley is a director of Aequitas Nexus, a not-for-profit organization empowering individuals through strengthened human rights protections and ethical AI tools to support vulnerable communities.
