Buildable — AI Real Estate Analysis
AI agent platform for Dubai real estate analysis. As Product Engineer, I own the path from client requirements to shipped platform — including an MCP analytics layer that turns usage data into roadmap tickets.
Tech Stack
Buildable - AI Real Estate Valuations
Overview
Developed the core financial valuation module for Buildable, an AI-powered platform that helps Dubai real estate professionals analyze deals in minutes, not hours.
The Challenge
Traditional real estate research and reporting is slow and manual:
- •Hours spent searching for comparable properties
- •Manual data cleaning and verification
- •Time-consuming report generation
- •Inconsistent methodology across teams
- •No real-time market intelligence
The Solution
Buildable transforms real estate analysis through AI and automation:
1. AI-Powered Valuations
- •Upload title deed → automatic property extraction
- •AI agent team built-in with real estate expertise
- •Professional reports generated in minutes
2. Market Intelligence
- •Verified DLD (Dubai Land Department) data
- •KPI benchmarks: prices, rental rates, yields, absorption
- •Outlier detection with AI flagging
- •Transaction history and trends
3. Smart Comparables
- •Instant matching by location, type, status, unit mix
- •Pre-filtered by square footage
- •No manual searching required
- •Export to Excel ready for reports
My Role
As Product Engineer, I own the path from high-level client requirements to shipped platform — assessing doability, scoping what to build (and what not to), making architecture calls, and delivering:
Key Achievements
Product Ownership- •Turn high-level client requirements into scoped, prioritized, shipped features
- •Built an MCP-based intelligence layer over product analytics, sales, and user-behavior data that suggests development tickets — usage data feeds the roadmap directly
- •Own prioritization trade-offs between client asks, platform direction, and engineering cost
- •Built real-time valuation engine
- •Integrated 5+ financial data providers
- •Improved calculation accuracy by 25%
- •AI agent system replacing ~6 hours of manual analysis groundwork per report
- •Agent memory with vector search (Qdrant), multi-LLM routing
- •Automated market reports, feasibility studies, and valuations
Technology Stack
Frontend: Next.js 15, React 19 (Bun), TypeScript, D3.js Backend: Python (FastAPI, Celery), PostgreSQL + PostGIS (SQLAlchemy), Redis, RabbitMQ AI/Agents: Multi-LLM routing, Qdrant vector memory, MCP, custom agent workflows Data & Search: DLD (Dubai Land Department) data, Typesense, multiple financial providers Infra: Docker, self-hosted Supabase (auth/storage), StripeImpact
- •2+ hours saved per valuation report
- •5+ data sources integrated seamlessly
- •25% improvement in calculation accuracy
- •Automated workflows for brokers, investors, developers
Who It's For
- •Brokers - Quick deal evaluation
- •Investors - Market intelligence
- •Developers - Feasibility analysis
- •Asset Managers - Portfolio valuation
- •Valuers - Professional reports
- •Consultants - Data-driven insights
The Future of PropTech
Buildable proves that AI can transform traditional industries. What used to take hours now takes minutes, with better accuracy and consistency.
---
*Current role: Product Engineer, Sep 2025 – Present*
