Job Description

Experience : 2-5 Years

Key Responsibilities

• Build and test extraction components
• Write and refine prompts and classification logic.
• Build parsers for different attachment formats, including PDFs, scanned documents, images, and email threads.
• Prepare and label unstructured documents for testing, and maintain evaluation datasets.
• Run accuracy tests, track results against confidence thresholds, and report findings to the team.
• Help implement review-queue flags for low-confidence, missing-information, or mismatched items.
• Support integration work with email inboxes and workflow queues, and help troubleshoot issues during testing and deployment.
• Write clean, documented code and unit tests.

Required Skills

• Good Python skills and understanding of core CS concepts (data structures, APIs, JSON, REST).
• Understanding of how LLMs work, and familiarity with prompt engineering through projects, internships, or coursework.
• Basic knowledge of text extraction from PDFs and images.
• Understanding of classification and metrics such as accuracy, precision, and recall.
• Comfort with Git and the command line.
• Attention to detail, since small extraction errors can send a document to the wrong channel.

Nice to Have

• Projects involving document processing, OCR, invoice or form extraction, or email automation (a GitHub link is a plus).
• Exposure to Azure, AWS, or GCP AI services.
• Basic SQL and experience working with structured reference data.
• Interest in or exposure to the insurance or financial services domain.
• Degree in Computer Science, IT, Data Science, or a related field.