There are many dating sites and apps in the world. From what i've seen, these apps tend to match show users pictures of people who are close to them. Then i thought, what if there is a system that gathers the pictures of the user's previous likes and figure out the type of face the user likes so that the system can display pictures that the user will most probably find attractive.
To build the system i used web technologies, Machine Learning(neural networks), face encoding and detection Libraries(dlib). The client side contains we apps to rate pictures and display predicitions(recommendations). It's written in html,css,javascript. The server side trains neural net models, organizes data and make predictions. it's written in python.
In this project, I've learned how to gather and organize data, integrate machine learning in a web app and be able to combine all processes into one product. The project took 2 weeks to complete.