Building the best way to pay!

Building the architecture for the best way to pay!

Course assistant for CSCI-UA.0467 - Applied Internet Technology at NYU Courant

Worked in team Video Understanding and on project Generic Watermark Detection.Built a pipeline to create a synthetic dataset for any watermark and train a deep learning-based image classifier to detect watermarks in Meta’s video stream data (Instagram stories, reels, videos).

Multidimensional Analysis of Trust in News Articles: Developed system to help users analyze news articles by examining if the content published by a media house is biased compared to the trusted representation. System developed using NLP, Graph-theoretic techniques, and Latent Dirichlet Allocation model. Work accepted for publication at the 34th Conference of Artificial Intelligence (AAAI) 2020.Feedback-based Keyphrase extraction from Unstructured Text Documents: Developed a framework for keyword extraction that enables customization via non-expert feedback in the form of natural language using NLP and machine learning. Work accepted for demo presentation at the 12th International Conference on COMmunication Systems and NETworkS(COMSNETS) 2020.

Worked on RoboTutor, XPRIZE Global Learning Finalist, an open-source Android tablet app that enables children aged 7-10 with little or no access to schools to learn basic reading, writing, and arithmetic without adult assistance.

Autonomous American Sign Language Translation: Developed a framework for generating written language from American Sign Language video by using state of the art Natural Language Processing and Image Processing techniques. Used a novel approach by treating Sign Language as a separate language rather than treating it as a gesture-based form of the spoken language. Curated an annotated dataset by scraping Youtube captions and videos.

Worked on the paper - "Applying Game Theoretic Approaches to Social Network Analysis" with professor Dr. Dolly Sharma.

Used the actual data set of the company and predicted whether a customer will claim insurance using predictive modeling and if so, how much, using regression techniques. Used machine learning models like KNN, Random Forest, Linear Regression in Python