Shirin Salehi received her PhD in Electrical Engineering from IUT, Iran, in collaboration with UPC, Spain. Since 2022, she has been a postdoc researcher at RWTH Aachen University, Germany, where she explores how to make modern AI models faster, lighter, and more sustainable. Her habilitation topic focuses on resource-efficient deep learning and large AI models for edge intelligence, with the goal of reducing the cost of training and inference while maintaining high performance. By rethinking how data, memory, models, and computation are used across fine-tuning, pre-training, and inference, we aim to develop techniques that enable powerful AI systems to run efficiently on edge devices. This talk will highlight why resource efficiency is becoming essential for the future of scalable and accessible AI.