It would be great to have a completely free solution that avoids Acrobat, but I've not seen one yet. The pdf's produced by djvu2pdf are ~10X larger than the original DjVu file, and I've only seen Acrobat's optimization able to get them down below the original file size. This is a significant issue for any size library.
Another track that Acrobat will do for the Kindle in this context is to split two-sided book scans into separate left/right pages for easy Kindle reading. The workflow is:
0. djvu2pdf, then save as optimized (as above).
1. Pages view, Right-click->Crop Pages. Crop the left, then the right side of the pages and save in separate pdf files.
2. Use the basic version of pdfsam
to merge the alternate left and right pages.
3. Acrobat OCR.
The resulting pdfs are typically the same size as the original DjVu files, but noticeably cleaned up, text searchable, and easy to read on a Kindle.