Building Sustainable Intelligence: Generative AI at scale

About the project

About the Project This project aims to pioneer advancements in the efficiency and scalability of Generative AI Models, focusing on achieving low computational costs at scale without compromising performance. As Generative models become increasingly vital to a wide range of applications, from generating images to generating videos and audio, their computational demand and the time required for training and inference have escalated. Their scalability is also limited, often resulting in short and inconsistent content generation. This research seeks to address these challenges by developing innovative techniques for efficiency and scalability, including architectural innovations, algorithmic improvements, and system-level optimizations that could enable deployment across broader scenarios of computing environments, from high-end servers to consumer-level devices. The overall objectives of your research will be: To develop efficient and scalable novel architectures and ML algorithms for training and inference. To create system-specific approaches and tools to monitor resource consumption, perform real time inference and dynamic resource allocation. You must have a UK 2:1 degree or its international equivalent. This opportunity is only open to home students . EU students can apply with a resident’s status. Applications will be reviewed on a rolling basis until the position is filled. Please contact the advisor directly. You need to have: Strong foundation in ML, demonstrated through completed projects and relevant coursework in mathematics and ML-related subjects. Excellent programming skills Background in Systems/Embedded computing (Desirable) Published research papers at top conferences/ Q1 journals (Desirable)

Research areas