What about feature extraction library for PDF

[dantetemplar]

I love the code you write. And I don’t like the code that I saw in the pdf-to-text libraries (I've seen quite a lot). Moreover, one of the fastest - mupdf does not reveal the full potential of the hardware, as I think. Perhaps your expertise in performance, SIMD, and AI can influence the scope of PDF (or other document types) analysis.

Ready to help in any way I can.

[ashvardanian]

I agree, that PDF processing libraries are horrible, but it's too far from infrastructure workloads we are focusing on. No plans to tap into that domain for now.