AI tech company Crowdworks (CEO Wooseung Kim) announced on the 22nd that it will be supplying its data construction solution, ‘Workstage,’ to a large-scale artificial intelligence (AI) project being promoted by a major domestic commercial bank.
This project was launched to link the bank’s large amount of document data with a large-scale language model (LLM). Given the nature of the financial industry, where security and data quality are highly valued, Crowdworks’ stable technology and data processing capabilities were the main factors in selecting the supplier.
Crowdworks designs a data pipeline that can systematically produce search and learning data through projects, and builds a learning data management system that can continuously supply high-quality 'AI Ready Data'. The system is an integrated solution that supports systematic collection, processing, inspection, storage, and version management of unstructured data, and enables internal assetization of the company.
The 'Workstage' being supplied this time is an on-premise-based data construction platform developed by Crowdworks, and its special feature is that it can safely process sensitive data in-house without external leakage. It is designed to efficiently perform the entire process from data collection to processing and inspection without separate coding knowledge.
In particular, it includes a 'Markdown Editor' function that allows you to easily extract information within a document, along with OCR and LLM-based automation functions. By simply specifying the box area of the document, text is automatically extracted, and spelling, spacing, table formats, etc. can be easily corrected.
Crowdworks CEO Woo Seung Kim said, “This project is the foundation for introducing AI into the financial sector, and securing high-quality data is the most important task,” and “Workstage is a solution that supports companies to secure learning data safely and systematically, and we will continue to expand into various industries in the future.”
- See more related articles
You must be logged in to post a comment.