Work/FORGE

FORGE

Describe your product, get a deployable landing page in 30 seconds. End-to-end AI-native product design.

Role

Founding Engineer — Frontend & Product

Year

2026–Present

The Problem

Building landing pages from scratch is slow. Existing website builders are bloated, and most AI builders produce generic, uninspired slop. Founders need high-end design that compiles to clean, perfectly responsive React code—instantly.

What I Built

LLM Generation Engine

A custom structured output pipeline utilizing OpenAI to synthesize raw text descriptions into a full JSON representation of the page layout, copy, and components.

OpenAI API · TypeScript

Real-time Live Preview

An entirely custom virtual DOM renderer that maps the generated JSON directly into a live, interactive React tree without reloading or deploying.

React · Webpack

Next.js / Tailwind Compiler

When the user is satisfied with the preview, the system compiles the layout to a production-ready Next.js project with Tailwind CSS utility classes and provides a downloadable ZIP.

Next.js · Tailwind

AI Design System Constraints

Implemented strict prompt engineering to force the AI to adhere to premium design system tokens (e.g. limiting colors to a curated monochromatic palette and enforcing high-contrast typography).

Design Systems · Prompt Engineering

The Stack

FrontendNext.js, TypeScript, Tailwind CSS, Framer Motion
Backend / EdgeVercel Edge Functions, PostgreSQL
AI PipelineOpenAI API (GPT-4o), LangChain
View Live SiteBack to All Work