Valco AI — A Real Estate Intelligence Platform for Dubai

Valco AI is a real estate intelligence platform I designed and built for Valco Properties in 2024 and 2025, as its Principal AI Engineer/Architect. It turns the Dubai Land Department’s open property data into market intelligence, and gives the company’s staff a conversational assistant that recommends the best options for their clients.

The challenge

Dubai publishes its property market as open data. The Dubai Land Department releases its records of sales, mortgages and gifts, of registered units and of rent contracts, and updates them daily.

The data is public, but it isn’t ready to use. It arrives as large CSV files, with names in Arabic and English, text where a database wants a key, empty values written as the word “null”, and dates written day first.

The questions that matter are also not the ones the files answer. A file lists what was sold. Valco’s staff need to tell a client where the market is heading, which properties suit them and what is worth buying.

What I built

• Data pipelines. Automated ETL pipelines ingest the Land Department’s CSV files, cleanse and enrich them, and load them into SQL Server and MongoDB. Each file is read as a stream, one line at a time, so its size doesn’t matter. Cleaning turns empty values and the word “null” into real nulls and puts every date into one format, and the cleaned files are bulk-loaded into staging tables.
• A model of the market. Set-based SQL turns the staging tables into a model built for analysis. Lookup tables hold the areas, projects, master projects, buildings, property types and room counts, and the nearest landmark, metro station and mall. Fact tables hold the transactions, the units and the rent contracts, and refer to those lookups. Sales and rents share the same lookups, so a building’s sale prices and its rents can be set side by side.
• Models. Time-series analysis and forecasting, clustering, anomaly detection and predictive scoring models run on the cleaned data. They are written in Python on PyTorch, TensorFlow and Keras, and include recurrent neural networks (RNNs).
• Conversational recommendations. Valco’s staff talk to the platform in plain language. It is a recommendation system: it finds the best options for a client and answers with market insight in context.
• Reports for clients. The platform produces market analysis, segmentation and forecasting reports that staff give to their clients.
• The dashboard. An analytics dashboard built in Blazor presents the market intelligence and the investment scores.

My role

I was Valco Properties’ Principal AI Engineer/Architect in 2024 and 2025. I designed Valco AI and built it end to end: the ETL pipelines, the data model, the forecasting and scoring models, the conversational layer and the dashboard.

Outcome

• Delivered to Valco Properties: the data pipelines, the models, the conversational recommendation system and the dashboard

Stack

C# · .NET · SQL Server · MongoDB · Blazor · Python · PyTorch · TensorFlow · Keras