An open source web application to collaboratively build and visualize neural networks with a simple drag-and-drop interface.

A simple extention of VAEGANs that employs label conditioning in additional to produce high resolution and high quality generated images. By adding an auxiliary classifier to the discriminator of a VAEGAN, the discriminator produces not only a probability distribution over sources but also probability distribution over the class labels.

EvalAI is an open source web application that helps researchers, students and data-scientists to create, collaborate and participate in various AI challenges organized round the globe.

A content recommendation system built using an ensemble of LightGBM and Deep + Shallow Neural Network for personalized content delivery. Model uses custom user & content features to rank the content based on past interactions on Glance app.

A Siamese Neu​ral Network model designed to classify whale species using whale fluke images. Problem involved classifying whale fluke to one of 5005 categories with imbalanced data.

Built a Squeeze-and-Excitation Network to build a multi-label classification model (1000+ classes). Trained using Focal loss & threshold search from pruned search space figured out by heuristics gathered from data analysis.

A bot designed to play the game FlappyBird using neural network, the bot learns to play the game and gradually improving using genetic algorithm.

A deep neural network designed for identification of thorax diseases from chest x-ray images using transfer learning on VGG-16.

A network traffic classification tool designed using libpcap and python which uses deep packet inspection and machine learning techniques to classify network traffic.