Time: 14:00 on 10/4/2018
Place: room 816, Ta Quang Buu Library, HUST
Presenter: Trinh Xuan Tuan, CEO and founder of NextSmarty
Title: Real-time Recommendation System for E-Commerce
Abstract:
In this seminar, we would like to talk about (1) the application of recommendation system in e-commerce for small-to-medium companies, and (2) the crucial importance of engineering aspect in deploying into production environment. In the first part of the talk, we present the unique characteristics of recommendation system in e-commerce setting, and the common pitfall of traditional method when approaching this problem. After that, we describe how we exploited these characteristics to design a novel algorithm, which achieved state-of-the-art performance on both offline- and online-assessments for e-commerce recommendation. The results were published in academic-track at RecSys’17, a prestigious conference in Recommendation System. We have also successfully deployed our solutions for clients worldwide, in which important business metrics such as conversion rate increases by 25% on average, resulting to around 50% of sales being generated from our recommendation engine. The road to delivering research work to the industry has taught us many valuable lessons, which forms the final part of our talk about real-time recommendation system at large scale in production environment.
Mr. Tuan is the CEO and founder of NextSmarty, a startup whose mission is to provide recommendation services by applying the state of the art in Artificial Intelligence. Before founding NextSmarty, he worked at DeepVu, a Silicon Valley-based startup, where he designed recommendation algorithms for American Express.
Tuan holds a M.Sc. in Computer Science from Uppsala University. He had also interned at Machine Learning Lab, Department of Computer Science, Copenhagen, where his focus work was on designing online learning algorithms for multi-class Support Vector Machines.