Smart RFQ

Project overview

Smart RFQ automates RFQ file generation for a global engineering services company. I worked on the project at QUIBIQ GmbH.

System workflow

1. Document access: Retrieve source documents from SharePoint.
2. Parsing and storage: Parse PDF and Word documents, then store the processed data in a database.
3. Keyword-based matching: Extract keywords using TF-IDF weights and compare texts using those keywords.
4. Efficient similarity search: Use an inverted index for embedding-based similarity to reduce runtime.

Deployment and technologies

The system was deployed using Azure Functions and Durable Functions. The workflow integrated SharePoint, PDF and Word document parsing, TF-IDF keyword extraction, database storage, and indexed embedding-based similarity search.