Image Captioning using deep learning. Currently, read and understood the papers on GAN, DCGAN, Conditional GAN, traditional encoder-decoder with attention approach, SeqGAN and Self Critical Sequence Training (SCST). Working on implementing image captioning with a combination of SeqGAN and SCST algorithms using PyTorch.

• Developed an unsupervised framework for ad-hoc retrieval of documents (approximately 150k) given queries

• Designed Autoencoder based model to perform sentence similarity to identify when user feedback should be considered

• Used Google’s Universal Encoder Model followed by clustering to predict context specific similarity

• Productionized the solution by deploying it on Flask app, researching on Big Data services which adhere to the requirements