Python Developer for Vector PDF and CAD Drawing Data Extraction

I’m looking for an experienced Python developer to help build a data-extraction pipeline for technical engineering drawings. The drawings are CAD-exported vector PDFs, and the goal is to turn them into clean, structured JSON that meets clear acceptance criteria. You’ll extract text, line and shape geometry, and CAD layer information from the drawings. You’ll also parse instructions from Word documents and email text, and extract text and tables from text-based technical reports. Another part of the work is grouping and counting drawing elements on each sheet against a set of defined rules, for example identifying how many of a given feature appear on a page. You should write strong, production-quality Python and have hands-on experience with vector PDF internals, using libraries such as PyMuPDF, pdfplumber or pdfminer. You should be comfortable reading drawings, text with coordinates, and optional content layers from a PDF. Experience with geometry and spatial processing, such as Shapely, is important, because much of the work involves proximity, grouping and bounding boxes. You should also be comfortable defining and validating structured data with JSON schemas and Pydantic. It would be a strong plus if you have worked with CAD or GIS data (such as AutoCAD, DXF files or ezdxf), or with engineering, construction or AEC drawings. Experience with computer vision for symbol detection, such as OpenCV or template matching, would also help. Familiarity with Microsoft Azure services such as Document Intelligence, Functions and Blob Storage is a bonus.