What we're about
Upcoming events (1)
How We Optimised Hero Images using Multi-Armed Bandit Algorithms with EPAM.
Join DSF in July for our Customer, Marketing and Growth Talks. Speakers from this sector share their stories, projects, joys, trials, and tribulations over the course of July 2021. Come and listen to these amazing companies and ask questions to learn more about this growing industry.
Ticket Allocation Process:
Registering here guarantees you a ticket for the Data Science Festival Virtual Event on July 29th , 2021. Please ensure to add this session to your schedule in order to receive the joining URL links.
Registration Link: https://growth.datasciencefestival.com/talks/dsf-session-5-epam/
Summary: How We Optimised Hero Images using Multi-Armed Bandit Algorithms. In this masterclass, we introduce the multi-armed bandit (MAB) optimization campaign we ran on Hotels.com™. The objective was to find the best possible main image (also called ‘hero image’) for the properties on our website whilst minimizing the impact of exploring different candidates. We will go through some of the main considerations and outcomes of this project.
Gyula Magyar - Lead Software Engineer - EPAM
Bio: Gyula is a Lead Software Engineer at EPAM who works with Expedia to help them build a multi-armed bandit platform. He specialises in data streaming with Apache Flink and Java backend development.
Humiun Miah - Product Manager - Machine learning & Personalisation - Expedia
Bio: Humiun is a product manager leading personalisation products at Expedia. He has successfully led product teams in driving human centred AI products and has been the leading force in instilling AI product principles at Expedia. Humiun has driven innovative products using machine learning in imagery, recommendations & ranking.
Fedor Parfenov - Lead Machine Learning Scientist - Expedia
Bio: Fedor is a Lead Machine Learning Scientist at Expedia who specialises in using contextual multi-armed bandits to optimise web page layouts, content and backend services. Previously he has worked on a/b test methodology, bidding and anomaly detection.