Ali Abdel Aal

A Python package that supports SOAP interface to communicate with the Microsoft ATKS, enabling the Python community to use the ATKS tool in their natural language processing based apps.

A dashboard for any custom query with simple analysis tools.

A project that had the purpose of targeting customers based on their social media feed using natural language understanding.

An Arabic Egyptian voice assistant with a chat-bot extension on Facebook.

My graduation project is a virtual assistant that interacts with you using your voice commands that are in Egyptian Arabic.

A variety of tools were integrated into the system to enable high accuracy and performance.

Mwaslaty is an Android application that helps the user find the shortest path and minimum time and cost to reach their destination. It was designed to minimize the effort of the user when trying to get directions to a new destination.

My role was to clean the data that is processed by the algorithm.

Using Python, I was able to achieve a good level of data cleaning, provided project-ready data that included translations and finding matched words in Arabic, and uploaded the data to a database.

The system monitors the car's balance. When the car drives over a manhole or bump, the system detects the change and sends the location from a Skylab GPS. The modules used in the system include the accelerometer, gyroscope, ultrasonic, and GPS. The system was built using an Arduino environment.

a classification model that classify whether the EURUSD stock exchange will go up or down next day based on historical data.

A tutorial on how to deploy a Flask app on Heroku server.

learn how to make a telegram bot and deploy it on a Heroku server, the bot will be up 24/7, and you can inject your own brain and responses into it.

Learn how to compress and decompress images using Keras.

The world's first Arabic voice assistant. My responsibility included delivering new features to Salma, from modeling the idea and doing a literature review on the idea to delivering it fully coded and functional.
Together with a wonderful team, we were able to deliver the first Arabic voice assistant.

Detects plagiarism in Arabic documents using distance function and machine learning.

The system model the relation between two articles by calculating multiple distance metrics and a machine learning model is trained on these metrics to classify whether there is a plagiarism or not.

Verify a tweet’s author by constructing a vector representation for the user by averaging the tweet's vectors that were created using word2vec.

The vectors are weighted with TF-IDF to enable sort of attention to the vectors that are important and a classifier is trained on the weighted vectors to predict the user.

Match user input with the correct semi-generated response for a chat-bot that is mainly focused on restaurants, based on TF-IDF and other similarity calculations.

The bot was provided as an API that you can send new data to it to train on and sample of the required responses.