At Julfar Consultancy Services, I built request-specific LLM agents for the RAK Labor Court that calculate legal claims, with amount accuracy reaching 93–100%. I also helped raise early-exit detection to 98.8%.
I built verdict generation and Arabic document extraction for contracts, payment proofs, and lawsuit filings, validating the work on 300+ real cases. Verdict drafts reach judges through the court’s SAP system in about five minutes.
For invoice verification, I refined extraction prompts and SQL queries to raise validation accuracy from 83% to 99.4%. I also built quotation matching and an LLM-as-a-judge framework to benchmark prompt versions.
As an independent Freelance Software & AI Engineer, I built a grant-tracking CRM using Gemini to turn funder web pages into schema-validated records, with human approval before saving. My personal Egyptian Labor-Law RAG Assistant provides bilingual, article-cited legal Q&A over Egypt’s Labor Law No. 14 of 2025.

