At Corelia, I designed a multilingual question-and-answer system for French documents that retrieves text and embedded images. I also developed LLM-based contract extraction and text-parsing pipelines that produce structured data.
For Monsif, my graduation project, I built an Arabic legal document analysis system with summarization that reduced reading time by 40%, and trained models to predict judgments across Egyptian criminal cases. I co-authored a paper on the work, published in Springer Proceedings and presented at AMLTA 2025.
I’ve also built local Arabic retrieval and chatbot applications, a multi-agent deep research system, and production-style RAG backends. Alongside this work, I taught Python, Java, and applied machine learning to more than 50 students across three programs.
