Description

Responsibilities

• • Build the Python-based AI agent/module for document analysis.
• • Support transient parsing of financial documents without retaining raw documents in PCC systems.
• • Extract derived, compliance-safe outputs such as recurring income patterns, estimated balance ranges, income categories, and transaction flags.
• • Exclude sensitive raw details such as account numbers, routing numbers, full transaction lists, exact transaction descriptions, check images, and unnecessary PII.
• • Implement prompt orchestration, document parsing, validation, confidence scoring, and error handling.
• • Integrate AI outputs with the Java/BFF layer through secure APIs.
• • Add observability for LLM calls, document parsing quality, failures, latency, and cost.
• • Support human-in-the-loop review patterns so AI findings assist the BOM/Medicaid Coordinator without making eligibility determinations.

Required skills

• • 6+ years of experience Strong Python engineering
• • Strong Python engineering.
• • Experience with LLM application development.
• • Experience with document parsing and structured extraction.
• • Fast API or similar API framework.
• • Experience working with PDFs, OCR/document intelligence, tables, financial statements, or semi-structured documents.
• • Strong security/privacy mindset around sensitive financial and healthcare data.
• • Experience designing reliable API integrations between AI services and product applications.
• Nice to have
• • Lang Chain, LangGraph, Azure OpenAI/OpenAI, document intelligence, or similar tooling.
• • AI evaluation frameworks.
• • Human-in-the-loop AI workflows.
• • Healthcare, fintech, or compliance-heavy AI experience.
• • Experience with non-retention / transient-processing architectures

Key Skills

Education

Any Graduate