AGMNT — E-commerce Recommendation System
Overview
Built the core ML infrastructure and full-stack platform for AGMNT, an e-commerce platform connecting Asian brands to Western markets. Serving 10K+ users across 15+ brand partners.
Technical Work
Recommendation engine — TensorFlow collaborative filtering system analyzing user behavior and product signals, improving conversion rates by 28% and average order value by 35%.
Full-stack platform — Node.js backend, React frontend, MongoDB. Built for scale from the start.
A/B testing framework — Processes 1M+ events daily. Used to drive all major product decisions including recommendation algorithm changes and UI experiments.
Role
CTO. Responsible for all technical architecture and engineering decisions.