Darin Tanaka

Led an end-to-end hybrid AI recommendation engine, lifting conversion 25% across 40M+ customers via 6-week A/B testing and production deployment on SageMaker/Redis. Built real-time NLP trend analysis and a retrieval-augmented virtual stylist, cutting trend-to-product turnaround 60% and reducing support tickets 30% while mentoring two engineers to promotion.

Reduced manual inventory planning time by 50% across 3,000+ brands by leading the Cloud Assortments Platform as tech lead for an 8-person team. Architected scalable GraphQL APIs for 500,000+ retail partners with 99.9% uptime and delivered Vertical Assortments and Targets to improve department goal visibility by 35%.

Built the Behold iOS app from the ground up as sole lead engineer, improving user retention by 40% and recommendation accuracy by 35% with a hybrid collaborative filtering + embeddings model and SageMaker serving. Reduced new brand integration time by 70% via a standardized zero-inventory drop-ship backend across 100+ brand APIs and increased weekly active user re-engagement by 45% using Firebase

Architected a GraphQL federation server to unify Nordstrom and Nordstrom Rack customer identity systems, supporting the 2018 Nordy Club loyalty program launch for millions of customers. Reduced return processing time by 30% across 240+ retail locations by building a state-machine-driven in-store return UI and improved reverse logistics throughput by 25% with a React Native employee shipping/wareho

Built an ad content filtering tool (AngularJS + Java Spring MVC) that reduced manual ad review time by 70% across 60M+ monthly visitors, protecting $1B+ in annual advertising revenue. Architected YP’s first A/B testing platform to enable data-driven ad experiments, driving an average 18% CTR improvement on YP.com.

Earned a Bachelor of Science in Computer Engineering from Purdue University in West Lafayette, Indiana.