Sunday, October 15, 2023

I developed a medical AI assistant app with a Flutter front-end and a Python back-end server using Flask and REST APIs. Utilizing local CRUD authorization and Firebase authentication, I created a seamless user experience, additionally connecting the backend using Flutter HTTP protocol and REST API services with the NewsNow API. 

MedBot HomePage

The app implemented a medical NLP diagnosis model based on user symptoms using TensorFlow. For the training and testing dataset, I utilized a public WebMD dataset for customized symptoms and Pandas, matplotlib, and other data visualization tools for the dataset analysis. The model achieved a 98% accuracy for training using LSTM units in NLP model and I additionally incorporated HuggingFace API for model training. Finally, I connected the model to my front-end Flutter app using Flask backend server with GCP deployment. 

Check out the project at: