HomePraise — From raw data to 92% accurate property valuations.
Real estate buyers and agents had no reliable, transparent way to get AI-powered property valuations. Existing tools were opaque, monolithic, and priced out of reach for independent agents. There was no platform combining geospatial intelligence, ML-driven pricing, and subscription-ready billing in one product.
Built a full-stack real estate SaaS from scratch. Trained TensorFlow ML models on real property datasets — feature engineering across location, size, amenities, and market trends — achieving 92% forecasting accuracy. Integrated LLM APIs for automated property description generation and natural language loan estimation. Developed geospatial search using Google Maps API with custom overlays. Implemented Stripe for subscription-based access tiers. Designed a microservices architecture with multi-tenant PostgreSQL isolation supporting 500+ simultaneous property listings securely.
92% ML forecasting accuracy on property valuations. 500+ listings managed with full multi-tenant isolation. LLM-powered content generation and loan estimator live. Stripe billing automated end-to-end. Full-stack from model training to deployed production UI.
