Data & AI
AI-Powered Customer Intelligence
Turning terabytes of data into measurable revenue.
ClientE-commerce Scale-up
Duration12 months
Year2024
Stack
PythonTensorFlowGCPBigQueryMLOps
Overview
Built a machine learning pipeline processing terabytes of customer data to generate predictive insights and personalise user experiences.
The Challenge
The client had years of rich customer behavioural data stored in siloed warehouses with no mechanism to derive actionable intelligence. Marketing campaigns were generic, churn was high, and lifetime value was stagnating.
Our Solution
We built an end-to-end ML pipeline on Google Cloud Platform — from data ingestion and cleaning in BigQuery to model training with TensorFlow. Propensity-to-churn and product recommendation models were deployed as real-time APIs, feeding directly into the client's CRM and e-commerce platform.
Key Results
- 123% reduction in customer churn within the first quarter
- 234% uplift in email campaign conversion rates
- 3Processing pipeline handles 2TB+ of event data daily
- 4Recommendation engine serves 4M+ personalised suggestions per day