🕷️️ Job Radar • SCRAPING

upwork.com 🟢 2026-05-12

🔹 [Target] Enhance LLM-based document data extraction accuracy and efficiency
👤 Client: 🇺🇸 USA Member since 2025-02-19
💰 Price: ****
🚩 Problem: Current auto-conversion process requires manual intervention for exceptions, limiting scalability.
📦 Existing: LLM models, existing mapping tables, daily standups with product and engineering teams.

Specifications:

[Target] Increase end-to-end conversion rate from current percentage to a higher level over six weeks
[Method] Review extraction exceptions, diagnose root causes, refine LLM instructions for specific document types
[UI/UX] Not specified
[Stack] Python (for scripting and data processing), LLM models, mapping tables
[Security] Ensure data privacy and security in handling sensitive documents during extraction process
[Format] JSON or CSV for structured output

Workflow:

1. Review extracted data against expected outputs to identify exceptions.
2. Diagnose root causes of exceptions by analyzing model behavior, document structure, and input data quality.
3. Refine LLM extraction instructions for specific document types and format variants based on findings.
4. Develop and maintain mapping tables that translate source-specific values to standardized outputs.
5. Identify systematic patterns in extraction failures and propose fixes or improvements.
6. Document all findings, recommend fixes, and collaborate with dev team to implement changes.
7. Participate in daily standups and weekly working sessions with product and engineering teams.