At Devsinc, I support a production document-processing backend handling 8,000+ uploads daily. I optimized AWS Lambda orchestration, reducing redundant function calls by 23% and lowering compute costs and timeout errors.
I built an AWS Textract pipeline to extract structured data from tax returns, insurance documents, and invoices, reducing manual review from 40% to 18%. I also built a RAG system with Qdrant and LangChain and integrated it into a live Django application.
At Muawin, I worked on vendor prediction and invoice analysis as an AI Engineer Intern. My projects include a tool for detecting and redacting sensitive information in PDFs and an evaluation platform for benchmarking document-processing pipelines.

