Shipped
Mizan AI
An AI legal assistant for Egyptian law. It helps people and legal professionals research the law, analyse court cases and process documents in plain language.

Why I built it
My family works in law. Growing up around it, I saw how much time goes into finding the right article, reading long judgments and preparing documents, and how hard Egyptian law is for ordinary people to understand. Mizan AI is my answer: legal knowledge that anyone can reach, in plain language.
What it does
- Legal research: ask a question in Arabic or English and get an answer grounded in the relevant Egyptian legal provisions.
- Court case analysis: upload a judgment or case file and get the key facts, arguments and outcome.
- Document processing: analyse contracts and legal documents and surface the clauses that matter.
- Guidance: context-aware explanations that help non-lawyers understand where they stand.
How it works
- Hybrid retrieval: BM25 keyword search combined with retrieval-augmented generation (RAG), so answers are drawn from a custom Egyptian legal knowledge base rather than guessed.
- LLM reasoning on top of the retrieved sources to produce clear, cited answers.
- Secure by design: user authentication and cloud deployment, built with privacy and security in mind from day one.
What I learned
Legal AI only earns trust if it is accurate and shows its sources. Building Mizan taught me to design retrieval pipelines carefully, test against real legal questions, and keep the interface simple enough for anyone to use.
Screenshots
More projects

Velocity Motors
Full-stack car dealership web app with browsing, booking system, contact functionality, and admin panel.

MedAI
A machine-learning web app that suggests possible conditions from the symptoms a user enters. Built with Python, Flask and scikit-learn as a learning project, not a medical tool.
Student Record Manager
Python / JSON
Student Record Manager
Built a modular console-based system in Python with dynamic GPA calculation (0–4.5), automated classification logic, and per-module pass/fail handling. Implemented strong input validation, JSON data persistence, statistical reporting, and automated report export, applying core concepts such as modular design, file I/O, and data validation.