Case Study: PropertyGuru automates its end-to-end ML pipeline with H2O Driverless AI

PropertyGuru saves 3 months from development to production with H2O.ai

PropertyGuru, a property technology company in Southeast Asia, faced a challenge scaling its machine learning processes. Its users upload thousands of property listing photos daily, and the company needed to automate several ML jobs simultaneously, including image moderation and churn prediction. Doing this work manually was not feasible, so they turned to vendor H2O.ai and its product H2O Driverless AI to automate its end-to-end pipeline.

The solution from H2O.ai involved using H2O Driverless AI to ingest data, perform automated visualization and feature engineering, and build models rapidly. PropertyGuru then deployed these models directly to AWS Lambda. This implementation saved an estimated three months from development to production, allowing the data science team to iterate faster and freeing them to focus on innovation.

PropertyGuru