Senior AI Engineer Needed – AI-Powered Tax Document Classification & Data Extraction (FastAPI)

I’m looking for an experienced Backend Engineer to build an AI-powered backend system that can process multiple types of tax documents, classify each document, and extract all filled fields and their corresponding values into structured JSON. The system must automatically identify the type of tax document uploaded, determine its structure, and extract all available information accurately. Scope of Work The backend should perform the following steps: 1.⁠ ⁠Document Upload: Accept tax documents through a single API endpoint. 2.⁠ ⁠Document Classification: Automatically identify the type of tax document being uploaded. 3.⁠ ⁠Data Extraction: Extract all filled fields and their corresponding values from the classified document, including relevant dates, names, addresses, tax identification numbers, financial figures, checkboxes, and other available information. 4.⁠ ⁠JSON Response: Return all extracted information in a well-structured JSON format, organized according to the document type. The system should handle multiple tax document types and support scanned documents as well as digitally generated PDFs, where applicable. Technical Requirements •⁠ ⁠FastAPI only for the backend API. •⁠ ⁠A single endpoint to handle document upload, classification, extraction, and JSON response. •⁠ ⁠AI/LLM-based document understanding and extraction. •⁠ ⁠OCR support for scanned documents, if required. •⁠ ⁠Structured and consistent JSON output. •⁠ ⁠Proper validation, error handling, and logging. •⁠ ⁠Maintainable, modular, and production-ready code. Expected API Workflow Input: A tax document uploaded as a PDF or supported image file. Processing: Document classification to Field extraction to JSON validation. Output: A JSON response containing the document type and all extracted field names and values. Deliverables •⁠ ⁠Fully functional FastAPI backend. •⁠ ⁠One API endpoint for the complete workflow. •⁠ ⁠Automatic document classification and field extraction. •⁠ ⁠Validated JSON responses. •⁠ ⁠Source code and setup instructions. •⁠ ⁠Basic testing with representative tax documents. Ideal Candidate You should have experience with: •⁠ ⁠FastAPI and Python backend development. •⁠ ⁠AI-powered document processing and information extraction. •⁠ ⁠OCR and PDF processing. •⁠ ⁠LLMs, structured outputs, and JSON validation. •⁠ ⁠Building reliable document-processing pipelines. When applying, please briefly describe your proposed approach, the AI/OCR technologies you would use, and any relevant document extraction projects you have built. Note: This project is focused exclusively on backend development. No frontend or UI work is required.