1. Zippy

A priority inbox system that uses an ensemble of TF-IDF models to determine the importance of an email and a RNN classifier to determine if the email requires any action from the user. Based on these information, the system would classify the emails into Urgent, Important and General. With global and user level TF-IDF model, the system would provide personalized recommendations. Moreover, with online learning, the system would continuously learn and improve its predictions for more personalized results.

2. Simple Devanagari recognizer

A simple character recognizer for devanagari script using CNN implemented using Tensorflow JS on Angular 7.

3. ProHealth

A simple web application to make communication between doctors and patients more accessible. It allows patients to add their problems along with their medical history and interact with doctors to get basic diagnosis. Doctors can interact with the patients in real-time to diagnose and prescribe medicines. It uses Django Rest Framework for REST APIs, React in frontend and websockets for real-time communication.

4. Chirp

A basic social media web application using Django similar to Twitter. Users can post, like other post, repost their posts and interact in simple ways.