I am a PhD student advised by Dr. Jeff Phillips at the School of Computing in University of Utah. Currently my research involves word embeddings and their geometry and the pursuit to make them more robust, interpretable and fair. My research interests broadly encompass various modules of Machine Learning, NLP and Data Mining..

Before joining my PhD position in 2016, I completed my integrated BS - MS in Mathematics at the Indian Institute of Science Research and Education, Kolkata where I did research on graph theory and brain mapping with Dr. Anirban Banerjee .

Research Projects

Absolute Orientation of Word Embedding Alignment

We explore ways to align high dimensional word embeddings extracted by different methods and from different datasets with each other. We also look at the different applications of this technique. We can extend these methods to other sets of high dimensional vectors.

arXiv Paper

Attenuating Bias in Word Vectors

In this project, we explore and compare ways to detect and remove bias from word embeddings. We perform a uniform simple linear projection on embeddings to correct the bias and demonstrate how it works the best among other methods. We also devise two quantitative methods to uniformly adjudge the gender bias and quality of the embeddings. Finally, we use names as a powerful tool to detect the bias direction of different types of bias like gender biased or racial bias.

arXiv Paper


Third year PhD student in Computing (Data Management and Analysis Track) at the University of Utah.

Completed a dual BS - MS degree majoring in Mathematics at Indian Institute of Science Education ad Research, Kolkata.


Teaching Mentee : Data Mining

Spring 2018, University of Utah

Teaching Assistant : Mathematics II

Spring 2016, IISER Kolkata


Here's a downloadable version of my CV!