Document Scoring Solution Assistance Needed

Hi Everyone,

I have been having some really interesting conversations with some potential clients and have come across a use case which could be productised across a lot of industries.

The requirements: Create a tool that can review an application document to output a likelihood of an application being successful, as well as reasoning, and highlight areas for improvement. There should be a mechanism for the model to iterate on itself with each further application to ensure it is improving over time.

The artefacts:

Defined criteria for success
100 examples of a mix of successful/unsuccessful applications

Has anyone had any experience implementing a solution like this and would be willing to collaborate on a PoC? My initial research has led me to look into embeddings, vectors and pytorch/tensorflow but I do not have the experience for this one.

Thanks!

5 comments

Document Scoring Solution Assistance Needed