I built an ML powered real estate pricing estimation product end to end
From requirements and database design to a production ready frontend, backend APIs, and AWS deployment
The app helps users search homes by ZIP code, explore listings on an interactive map, and open a detail view that explains estimated value using charts and feature level signals

What it does

• Search and discovery

  • Search by ZIP code with a fast listing experience
  • Filter by price range, beds, baths, and property type
  • Browse results as cards and as pins on Google Maps

• Property detail experience

  • Image carousel and listing summary
  • Under valued vs over valued gauge with an estimated price readout
  • Utility cost trends rendered from a forecast dataset
  • Property sale history table
  • Nearest school distance table
  • Key feature signals and a comparable homes block based on nearest matching values

Architecture and implementation

• Frontend

  • React SPA with componentized UI
  • Tailwind CSS for design system styling and layout
  • Ant Design for structured layout primitives and form controls
  • Google Maps integration for map based discovery
  • Data visualizations for pricing and utility trends

• Backend and data

  • Node.js based API layer deployed as AWS Lambda behind API Gateway
  • DynamoDB tables for property features, sales history, and ZIP level scoring
  • Amazon S3 hosted datasets used by the frontend for utility cost charting

• Product and delivery

  • I owned the build end to end including database design, integration work, and deployment
  • Prototyped parts of the data workflow with Supabase during early iterations and carried those learnings into the final production design
  • Delivered within the agreed timeline with a strong focus on stability and UX polish

Skills and deliverables

• React
• Node.js
• AWS Lambda
• Amazon API Gateway
• Tailwind UI and Tailwind CSS
• Amazon S3
• Database design
• Deployment and release ownership

Screenshots