Founder-built SaaS for real estate: turn property photos and an address into polished, audience-specific listing copy — with semantic search across the portfolio.
Context
nestd.ai is a product I designed and shipped end-to-end: concept, UX, backend, frontend, and go-to-market basics. It is separate from my client work, but it is the clearest proof that I still build and ship — not only advise.
What I built
- Vision pipeline — per-room photo analysis (room type, features, style, condition) to replace manual data entry
- Location context — neighbourhood and amenity enrichment from a property address
- Multi-audience generation — tone, length, language, and audience tuned output
- Semantic search — tool-calling agent with vector embeddings for natural-language queries over listings
- Product surface — trial flow, dashboard, and enterprise-facing landing page
Stack & approach
TypeScript full-stack, modern AI APIs, vector search, and pragmatic SaaS delivery — small iterations, production-minded defaults, and features shaped around what agents and brokers actually need.
Links
- nestd.ai
- Product scheduling (not client work): schedule.nestd.ai