I am a DPhil (PhD) candidate at the University of Oxford under the supervision of Nic Lane. I am part of the Machine Learning Systems Group and also a member of Linacre College. My research focuses on designing deep neural network architectures that can run efficiently on constrained platforms. In particular, I am interested in designing hardware-aware optimizations for deep learning workloads enabling efficient inference when consuming data in the form of images, audio and graphs. I am also interested in exploring on-the-fly computations to ease the problem of data movement in setups with very limited memory. My up-to-date (Oct. 2020) resume can be found here.
I completed my B.Sc in Telecommunication Engineering at TECNUN, University of Navarra (Spain), in 2014 and a M.Sc in Computer Vision at Queen Mary University under the supervision of Prof. Andrea Cavallaro in 2015. Prior to starting my DPhil, I worked as a computer vision engineer at: Hawk-Eye (Sony), where I improved their multi-camera calibration and player tracking software; and at Holition, an award-winning creative studio where I co-designed and implemented augmented reality applications for the fashion retail sector.
I can also be reached at —> javier [dot] fernandezmarques [at] cs.ox.ac.uk