Developed metadata harvesters to extract metadata from library sites of US universities. Helped build a digital library with 500K thesis and dissertation information.

A Transformer-CRF-based citation parser model trained with synthesized data containing 1 billion citation strings (GIANT dataset). Evaluated the models with the CORA benchmark and achieved a 100% F1 score on key citation fields.
Thesis

Learn about your intake in one casual voice command! This software allows users to find food nutrition values from their natural language commands using NLP tools and techniques. Read my blog to learn more.

Explored the dataset using EDA to find out prominent factors for heart failure which later were used to build up the machine learning model for early prediction.

Play Flappy bird in your browser!

This extensions mimics browsers zooming functionality. Added benefit is you can increase or decrese the zoom by 1 level each time instead of 10.

When you use browser zoom, ever wondered if the smallest ones gets zoomed out first? It would be great for the people with visual impairment. Well this extension is for you..[In Progress]

Research project for lower visioned people to enable them creating table of content from a webpage, to allow focusing on desired part of the webpage later. [In Progress]

Infered the origin and traits of students who do poorly in their academics from Open University Learning Analytics dataset

Using Logistic regression, developed a model which can detect if a text is harmful or not. No library was used.