Demand is higher than ever to create and use Machine Learning models in production systems. However, community attention is usually focused on the process of building and training a model; with little acknowledgment given to how these models are then deployed and used by applications, and the supporting infrastructure required. I shall detail a case study of the processes and tooling used to successfully integrate ML into M&S’s Try Tuesday online personal styling service; from inception to deployment. This talk will be useful for anyone wishing to utilise ML models into production systems in a robust and scalable way.
Chris is a machine learning engineer with interests in Deep Learning, Game Theory and Systems Architecture design; having obtained a BSc Mathematics and MSc Computer Science from the University of York. Chris currently works for M&S’s Try Tuesday online personal styling service; encompassing projects such as fashion relevant visual search, customer segmentation and targeted recommendations. Outside of industry, Chris has authored award-winning research papers on the application of extortion strategies to the iterated poisoners dilemma, and utilising the Monte Carlo tree search algorithm to create the first successful opponent model for the game Liar’s Dice.