E-commerce
Hyper-Personalization Engine for Major E-commerce Retailer
Designed a real-time product recommendation engine using collaborative filtering and deep learning to boost average order value (AOV).
Completed: 10/5/2024
🎯 Project Scope
Model development, A/B testing framework deployment, and seamless integration into the existing microservices architecture (checkout, product pages).
⛰️ The Challenge
Generic recommendations led to poor user experience and lost conversion opportunities. Need for real-time response to user behavior.
🛠️ The Hanva Solution
Utilized a hybrid recommendation model, deployed on a GPU-enabled cloud platform, serving predictions via a high-throughput API.
Key Technologies & Architecture
- • Python
- • PyTorch
- • GCP Vertex AI
- • Redis
- • FastAPI