Delivering perfect personalised style recommendations to Thread's customers takes the efforts of many functions across the company: styling, merchandising, UX design, copy-writing and AI combine to produce our “Style AI”. I will outline how we have built a data-driven multi-discipline approach to Style AI, where data from ML recommendation systems is fed-back to inform which products we should stock, and guide the types of "looks" our stylists create. This leads to a virtuous circle of improving recommendations over time as we learn more about what our users want to see, and are able to further tailor our suggestions to them.
A one-time ML researcher of Bayesian regression and classification algorithms, Ed has now worked on machine-learning systems in a diverse range of industries: from algos powering hedge-fund trading strategies, to search-ad ranking and optimisation at Microsoft Research / Bing Ads, and real-time bidding (RTB) algorithms for display advertising at Adform. Ed is now Head of Applied Research at Thread, where he develops the AI to help Thread's stylists make perfect personalised clothing recommendations for men.