Supercharging Fashion Retail with AI

TL;DR

Joined SeeStone in November 2025 as part of the founding team. Building AI-powered forecasting to help fashion brands predict product demand and optimize inventory orders, eliminating dead stock and waste. Working full-stack on frontend, backend, AWS infrastructure, and core forecasting algorithms using Gemini and Amazon Bedrock.

The Problem

Fashion retail has a massive waste problem. Brands order inventory for new products without accurate demand forecasting, which leads to Dead inventory, Monetary loss and Waste management. The root cause is uncertainty. Brands don’t know how well a new product will perform, so they either over-order (waste) or under-order (lost sales). Both are expensive mistakes.

SeeStone uses AI to predict how well a new product will do before brands commit to large inventory orders. We’re building forecasting and similarity algorithms that analyze product attributes, historical data, and market trends to predict demand. Brands use those forecasts to make smarter inventory decisions, reducing waste and maximizing profitability.

My Role

I joined in November 2025 as part of the founding team, working full-stack across: Next.js and TypeScript for the main app, Angular for the admin dashboard, API development, database design with Postgres, AWS (S3, EC2, Lambda, API Gateway, CloudFront, CloudWatch), core forecasting and similarity algorithms using ML.

It’s early-stage startup mode—building fast, iterating constantly, wearing multiple hats. The challenge is making the AI accurate enough to be trusted by brands making million-dollar inventory decisions. We’re getting there.