I need an ultra-efficient solution that reliably shrinks very large PDF documents—up to about 2,500 pages—down to roughly 7 MB without crashing or corrupting the file. The compressed PDFs will be uploaded to a web portal, so the resulting file must open quickly in a browser and keep text selectable while images are reduced to a sensible, medium-quality resolution. You’re free to build the compressor as a standalone desktop app, a command-line script, or a lightweight web utility; what matters is that it runs on Windows and can be integrated into an automated workflow later. Popular libraries such as Ghostscript, qpdf, or Python-based solutions are fine as long as licensing permits commercial use. Please make sure the compressor: • Accepts PDF input only. • Consistently produces output ≤ 7 MB for a 2,500-page test file. • Preserves searchable text and bookmarks while applying medium-quality image down-sampling. • Finishes a full run in a reasonable time (minutes, not hours) on standard hardware. Deliverables: 1. Executable or script with clear usage instructions. 2. Source code and build steps. 3. A short report detailing the compression approach and any adjustable parameters so we can tweak quality-to-size ratios in future.