Client: Pan-Asia E-Commerce Platform
Tools: Python, TensorFlow, Vertex AI, BigQuery, Redis
Challenge:
Personalization was limited to rule-based engines, failing to drive conversions or cart retention.
Solution:
- Built a deep learning-based recommendation system using collaborative and content filtering models
- Used Vertex AI Pipelines for automated model training and deployment
- Integrated with BigQuery ML for real-time segmentation
- Cached results with Redis to reduce latency on high-traffic product pages
Outcome:
Lifted average cart size by 22% and improved click-through-rate (CTR) by 35% in the first quarter post-deployment.