AI Automation Developer for Sock Design Catalog Generator

Job Title: AI Automation Developer: 2D Design Board to PDF Catalog Pipeline Role Overview I am looking for a talented developer to build an automated "Design-to-Presentation" engine. We currently create 2D design boards for custom socks. We need a tool (ideally hosted on Replit) that uses AI/Computer Vision to extract individual designs from these boards and "spit them out" into a professional, multi-page PDF catalog. The Project The goal is to eliminate manual cropping and slide creation. You will build a pipeline that: Ingests a PDF design board (containing multiple sock views) and a brand logo. Extracts metadata like Design Numbers, Pantone colors, and construction details (e.g., "Dress, Crew, Woven, OSFM"). Segments each sock design (Front/Back views) into individual assets. Generates a branded presentation (like the attached ICPSR example) with a cover page, one page per design, and a contact footer. Technical Requirements AI/Vision: Experience with Gemini 1.5 Pro, GPT-4o, or OpenCV/YOLO for layout analysis and image segmentation. Language: Python is preferred. PDF/Graphics: Proficiency in PyMuPDF, Pillow, and PDF generation libraries like WeasyPrint or ReportLab. Environment: Ability to deploy and hand over the project via Replit. Design Eye: The final output must look like a high-end marketing presentation, not a raw data dump. Deliverables A functional web app or script where I can upload a PDF board and download a finished catalog. Clean, documented code that allows for template tweaks in the future. How to Apply Please share examples of previous automation or "Vision-to-Document" projects you’ve built. If you have experience specifically with automated brand catalogs or garment tech packs, that is a huge plus!