Document AI Engineer: LLMs, Extraction & Production Systems

Weekday 1; Chicago (IL); On-site; USD 140,000 - 190,000; Full time

Job summary

Weekday is seeking an experienced AI Engineer to design and build the intelligence layer across a document-to-return workflow. The role focuses on production-grade AI that transforms complex financial and tax documents into structured data and supports professionals in identifying missing information, inconsistencies and potential errors.

This hands-on role emphasizes document OCR, LLM-powered understanding, agentic systems, evaluation, and human-in-the-loop workflows to ensure accuracy,

Qualifications

• 3+ years of experience building production AI systems.
• Hands-on experience with document OCR, understanding, parsing and structured data extraction.
• Experience with PDFs, forms, scanned documents, financial documents, or other unstructured data.
• Strong knowledge of LLM-based document processing and AI techniques for extraction and reasoning.
• Experience designing evaluation frameworks for AI or ML systems.
• Knowledge of ground-truth labeling, scoring methodologies, benchmarking and statistics.
• Experience with agentic AI systems and human-in-the-loop workflows.
• Strong observability, reproducibility, model monitoring and production reliability.
• Proficiency in Python and building scalable production systems.
• Bias toward shipping, experimentation, measurement and continuous improvement.

Responsibilities

• Build production systems for document classification, OCR, parsing and structured data extraction.
• Process PDFs, scanned documents, tax forms, financial statements, receipts and other unstructured financial information.
• Design extraction workflows that preserve source context, handle ambiguity and route low-confidence results for human review.
• Develop LLM-powered document understanding and intelligent extraction capabilities.
• Build AI agents that analyze completed returns against source documents and relevant tax context.
• Identify missing information, inconsistencies, potential errors and present findings clearly for professional review.
• Design human-in-the-loop workflows with confidence signals, source citations, review controls and correction mechanisms.
• Build scalable evaluation frameworks for document-processing systems.
• Define evaluation datasets, ground-truth labels, scoring methodologies, benchmarks and statistical analysis approaches.
• Establish observability and feedback systems to measure extraction quality, model performance and user outcomes.
• Monitor production AI workflows and continuously improve accuracy, reliability and coverage.
• Identify high-effort manual steps where AI can add value.
• Collaborate with engineering and domain experts to translate requirements into reliable AI systems.
• Establish reproducible testing and evaluation processes for models, agents and extraction pipelines.
• Contribute to expanding AI capabilities across document processing, return review and professional workflows.

Skills

Python

Observability

Production AI

Experimentation

Problem solving

Tools

OCR

PDF processing

LLM tooling

Job description

Weekday is seeking an experienced AI Engineer to design and build the intelligence layer across a document-to-return workflow. The role focuses on production-grade AI that transforms complex financial and tax documents into structured data and supports professionals in identifying missing information, inconsistencies and potential errors.

This hands-on role emphasizes document OCR, LLM-powered understanding, agentic systems, evaluation, and human-in-the-loop workflows to ensure accuracy,