Case Study: a US e-commerce giant accelerates machine learning model deployment with Persistent Systems

US E-Commerce Giant Accelerates Machine Learning Model Deployment by 80%

The US E-Commerce Giant, a leading home furnishings retailer, faced significant challenges with its machine learning recommender systems. These systems were failing to meet accuracy and performance benchmarks, causing a 10% impact on business growth due to missed opportunities. Its internal teams struggled with data consolidation and language barriers in their coding, which slowed model deployment. They commissioned Persistent to transform and optimize their data pipelines on the Google Cloud Platform.

Persistent developed a centralized data platform that abstracted feature engineering and homogenized data processing languages. This solution provided the client's ML teams with ready-to-access, cleaned, and consolidated data, accelerating model deployment time by 80% and reducing latency by more than half. The programmatic handling of 1000 features allowed for highly personalized recommendations, which subsequently improved customer engagement by 10%.