Senior AI Engineer for Form-Filling Automation

We are building an AI workflow that completes medical forms using a structured medical record as the source of truth. We receive a long tail of PDFs with different layouts, but most request the same underlying information about diagnosis, treatment, limitations, and leave. Some PDFs contain fillable fields; others are scanned or visually structured documents. Today, our operations team completes these forms manually. Our goal is to generate reliable drafts so operators primarily review and approve them, with a target of approximately 15 forms per hour while maintaining or improving accuracy. We need someone who has built similar production systems and can help us identify and implement the fastest credible path to this outcome. Scope - Review our existing workflow, architecture, and AI form-filling system - Recommend the right approach for both known and unseen form formats - Map information from structured medical records into form questions - Complete fillable and non-fillable PDFs - Add validation, confidence scoring, source evidence, and human review - Define an evaluation system based on field accuracy and operator edits - Build and test an initial production-ready workflow Ideal Experience - Production AI document-processing or form-filling systems - OCR, document understanding, PDF manipulation, and structured extraction - LLM workflows with validation and human-in-the-loop review - Python and modern AI APIs - Experience handling sensitive healthcare data is a plus We are not looking for someone whose experience is limited to chatbots, basic RAG systems, or simple OCR extraction. When Applying Please briefly answer: - What is the most similar system you have built? - What was your specific role and what measurable results did it achieve? - How would you handle forms the system has never seen before? - What could you realistically prove within the first one to two weeks?