Rosetta.ai — AI Recommendation Engine
Owned the product lifecycle for an ML personalisation engine deployed across 1,000+ e-commerce clients — turning generic "bestseller" widgets into real-time, per-user recommendations.
Challenge
Clients relied on generic "bestseller" and "frequently bought together" widgets that ignored individual behaviour and style preference — causing poor discovery, low AOV and high product-page bounce.
My role
As Product Specialist & Domain Expert I owned the lifecycle: defining model requirements with data science, managing data infrastructure, designing the A/B testing framework, and coordinating client rollouts across 1,000+ merchants.
Outcome
A real-time engine reading browsing patterns, style affinity and purchase intent to surface hyper-personalised recommendations — 99% model accuracy and 43% faster response than the previous solution, driving dramatic lifts in core commerce metrics within a year.
