Creating and correcting customer-facing product relationships at global Amazon catalog scale (billions of products in tens of marketplaces) using a calibrated blend of machine learning and human labelling, Big Data technology (Apache Spark/Java) and our Typescript/React auditor labelling UI, alongside specialised Java services. I've led design and evolution for some of our services, provided design guidance and mentorship for engineers both inside my team and across the organisation and built and maintained workflows on our Spark data pipeline.

Devising tools to automate vendor negotiations at scale. I was heavily involved in retail stakeholder engagement, developing deep domain understanding and eliciting requirements by working on-site and remotely with US-based internal vendor management teams and built high-scale map-reduce jobs in Java/Apache Spark to perform data analysis and machine learning at scale.

A very small experimental prototyping team to deliver paid content using existing internal Java merchandising systems. Having identified available business opportunities through stakeholder engagement, we architected a solution integrating disparate existing systems, requiring onsite and remote engagement with US ads teams. Once the concept was proven, I led final product handover to US-based team.

Served high-visibility recommendations experiences on Amazon home pages and elsewhere across all regions, through a constellation of Java microservices. Blended solid operational excellence with fast-paced iteration and A/B testing to improve customer experience. Tight latency requirements drove detailed focus on issues like caching, performance optimization, memory management, garbage collection and networking.