Will Humphreys

My portrait

I'm a data scientist with a strong interest in machine learning and visual arts. For over 2 years, I've worked at Egress as a developer and data scientist, developing APIs and services on Azure infrastructure, as well as delivering data analysis and product insight. I have experience with stastical techniques, but I also like to apply unsupervised learning methods where possible.

I studied (BSc) Computer Science with a Year in Industry at the University of Sheffield. In July 2020, I gradauted with a first-class degree. My dissertation, Detecting Individuals in Multiple-Shot Video Sequences focused on person re-identification for televison. I made use of TensorFlow, OpenCV and dlib to implement a β-Convolutional Variational Autoencoder with a custom KL-annealing scheduler, intended for the unsupervised clustering of facial features.

As an artist and illustrator, I work with oils, inks, watercolours and a bunch of software. I've live-coded visuals at Algoraves using the Hydra library, which is what powers the header of this page. I've also done video for corporate events and gigs with a wide range of projection mappings using Resolume.

I was involved in the student hackathon scene, co-organising HackSheffield 3.0 as a member of the commitee and assisting in the organisation of HackSheffield 4.0. I attended quite a few hackathons, developing using a wide variety of technologies .