A machine-learning personalisation engine that scores products in real time — powering feed ranking, cart upsell and campaign segmentation for a multi-brand retailer.
Representative engagement — anonymised and illustrative. It describes the kind of problem we solve and the platform we build, not client-confidential performance data.
A multi-brand retailer had strong traffic but weak conversion. Product discovery was generic, recommendations were rule-based, and campaign targeting had low relevance across user segments.
We implemented a machine-learning personalization engine that scores products in real time based on behavior, intent, and inventory context. The engine powers home feed ranking, cart upsell, and campaign segmentation.
Real-time behavioural scoring replaces static rule-based recommendations
Personalised feed ranking makes product discovery relevant per shopper
Cart and campaign targeting draw on the same intent signals
Recommendation latency engineered low enough for live shopping sessions
Design and build of an AI-assisted claims processing platform that replaces a decade-old manual workflow with digital intake, automated data extraction and rule-based routing.
Product strategy, architecture and delivery of a multi-tenant property-management SaaS — taking a founding team from vision to a production platform with secure rent payments.
