Back to work

AI Commerce Platform / 2025

Zory AI

Zory AI is a multi-tenant AI commerce platform for furniture retailers, built to turn room photos, product catalogs, and store data into relevant product recommendations.

Timeline

18 months

Role

Technical lead, product architecture, full-stack engineering

Zory AI commerce dashboard showing visual search, furniture recommendations, store analytics, and product performance data.

2

major product phases launched

5

engineers led

100%

custom AI model deployed

(Case study)

The work behind Zory AI.

Challenge

Furniture shoppers often know the look they want before they know the product name. Zory AI needed to recognize room context, match it to merchant inventory, and still behave like a reliable SaaS platform for many stores at once.

Approach

01

Architected the Django and React SaaS foundation with tenant-aware stores, roles, inventory, analytics, and e-commerce integrations.

02

Trained production computer vision models with YOLOv12 and RF-DETR to recognize furniture, decor, and room context from images.

03

Connected image embeddings, vector search, and Pinecone recommendations so merchants could surface visually similar products from their own catalogs.

Outcome

Zory AI shipped through two major product phases with custom AI models in production, a working visual recommendation engine, and a scalable foundation for future room redesign and AI shopping workflows.

(Interface)

The product in practice.

Zory AI commerce dashboard showing visual search, furniture recommendations, store analytics, and product performance data.

Project type

AI Commerce Platform

Role

Technical lead, product architecture, full-stack engineering

Timeline

18 months

Core stack

PythonDjangoReactAWS ECSPineconeYOLOv12RF-DETRREST APIs

Next case

Gaming & AdTech Platform / 2024

TechWrap

View project