Project Overview
StyleAI, a fashion e‑commerce store, wanted to increase cross‑sell and upsell. Intelirate built a hybrid recommendation engine (collaborative + content‑based).
Services Provided
- User‑item interaction logging (clicks, purchases)
- Matrix factorisation (ALS) + item embeddings
- Real‑time serving via REST API
- A/B testing framework for recommendation strategies
- Personalised email recommendations
Results Achieved
- Average order value increased 35%
- Click‑through rate on recommendations 22% (vs 8% generic)
- Revenue uplift $4.5M in first year
- Customer return rate up 25%
Client Feedback
“StyleAI’s engine feels like a personal shopper. Intelirate turned our data into gold.”
Tech Stack
- Python, TensorFlow Recommenders
- Redis (online serving), PostgreSQL
- AWS SageMaker, Kubeflow