Description: I need someone to build a piece of software that allows engineers to search through a large library of technical documentation (around 300GB of PDFs, scanned manuals, and electrical diagrams) using natural language.
This software will be used by marine engineers to ask questions like:
“Where is the 24VDC power supplied from for Cabin 12 lighting?”
The tool should return accurate answers based on the content of the documents, ideally showing the source (e.g. file name and page). Some of the documents include wiring diagrams and line drawings, so it must also be able to interpret and return information from image-based content, not just text.
Key requirements: • Runs as a standalone Windows application • Has a simple, user-friendly chat interface • Can process and index large volumes of documents (PDFs, images, scanned content) • Understands and extracts data from both text and diagrams • Uses AI to return accurate answers to engineering queries • Stores all data locally and securely (only API requests for AI if needed)
This needs to be a real, usable software product — not a prototype or script — ideally something that could be installed via .exe or similar and used by non-technical crew onboard.
Please include relevant experience or past projects when applying.
Price is just there for posting job, please send realistic proposal.
