Description
Summary We are looking for a Python developer to build a paid prototype of a batch metadata processor for OCR’d historical publication PDFs. This is not a chatbot project and not a prompt-only project. We already have tested metadata rules and Custom GPT prompts. We now need a small, reliable prototype that turns those rules into a repeatable, validated processing workflow. The prototype should process a small test set of OCR’d PDFs and authority spreadsheets, create one page-level metadata row per PDF page, use the OpenAI API for page interpretation, validate final names and chapters against authority files, and export review-ready Excel/CSV files. Successful completion of this prototype will likely lead to the next phase: a fuller batch processor for several hundred OCR’d publication issues. The attachment contains a more detailed project description and has a tab containing screening questions that must be answered before being selected for the project.