Applied AI · Real estate
AI-assisted property discovery
I led architecture and engineering delivery for a real estate platform serving the Portuguese market.
MY CONTRIBUTION
What I worked on
Property search and assistants
I worked on LLM-powered assistants that interpret buyer intent and apply listing filters, with quality checks before release.
Listing data pipelines
I coordinated listing ingestion, asynchronous processing and deduplication, with review workflows for data-quality issues.
DESIGN & VALIDATION
One decision in detail
Engineering decision
Keep property filtering in the application. The assistant interprets the request and the listing search applies the actual constraints, so the response can be checked against the returned records.
Example acceptance check
“Show me two-bedroom homes near Porto under €350,000.”
Check that the location, bedroom count and price ceiling carry through to search. If no listings match, the assistant should say so and ask whether to widen a filter.
What the design enables
The workflow connects conversational search with structured listing data and gives reviewers a way to inspect where an incorrect match came from.
PROJECT INQUIRIES
What are you
building?
Tell me about the product, workflow or AI feature you want to build and where you need help.