Case Study: Prophia accelerates commercial real-estate document extraction with Hugging Face and AWS

Prophia - Customer Case Study

Prophia, a company in the commercial real estate sector, faced the challenge of manually extracting critical data from lengthy and complex lease documents. This process was both tedious and prone to errors, making it difficult for business users to leverage the information for portfolio management and asset valuation. To address this, they turned to the vendor Hugging Face, utilizing their models and services on AWS SageMaker to build a robust NLP solution.

The solution involved leveraging Hugging Face's Deep Learning containers on SageMaker to fine-tune models like LayoutLM, Roberta, and T5 for their specific needs. This allowed Prophia to automate the extraction of over 140 different data points from documents. The implementation with Hugging Face resulted in a highly efficient ML pipeline, leading to significant resource efficiency, cost savings, and a vastly accelerated process for transforming documents into business-ready data.

Prophia