Using ML to build a model from large texts

Preamble: I know only a smidge about ML--took & 'certificated' the Coursera ML course, but that's it.

I have a collection of text documents. Each document is roughly 100 pages of scientific text (the documents are research proposals). Each document is ranked by a review committee and the rankings determine if the research proposal is funded.

I would like to see if it's possible to use ML to build a model that can predict review committee rankings by training on older research proposals.

• Is there a particular ML method that is more/best suited to examining large texts in this way?
• Suggestions on where to look to see similar types of model building?
• General comments?

Thanks for any insight!