Understanding Customer Intent in the Home Improvement Domain
, Senior Data Scientist, Lowe's
, Director of Applied Research, Lowe's
Recommender systems are key in ecommerce, helping customers find the most relevant products in an information-overloaded environment. Given the diverse set of products and multifaceted user objectives in home improvement, capturing user intent in real time is pivotal for delivering accurate recommendations. Recent advances in sequential and transformer/attention-based architectures underscore the efficacy of next-item prediction tasks grounded on user intent. We'll delve into the unique challenges of real-time recommendation in the home improvement domain. We employ NVIDIA Merlin models to facilitate real-time recommendations for product recommendations.