Vipul Gupta

Vipul Gupta

Ph.D. Candidate in NLP lab at Pennsylvania State University

Research Scientist Internship at FAIR, Meta | 2024

NLP Research Scientist Summer Internship at Kitware | 2023

Research Intern at John Hopkins University | 2021

Research Assistant at University of British Columbia | 2021

Indian Institute of Technology(IIT) Delhi - B.Tech in Mathematics and Computing | 2018

I am a Ph.D. candidate in Computer Science at Pennsylvania State University, in the NLP lab under Prof. Rebecca Jane Passonneau My research focuses on Responsible AI and model evaluations, with a particular emphasis on large language models.

Currently, I'm contributing to the field of AI as a Research Scientist Intern at Meta Fundamental AI Research (FAIR) group. As part of Responsible AI Team under Adina Williams, I'm leading projects to identify reliability issues in benchmark datasets for Large Language Models and developing methodologies to enhance the robustness and efficiency of evaluation datasets. In the summer of 2023, I worked as an NLP Research Summer Intern at Kitware, contributing to the In The Moment project funded by DARPA. I focused on application of large language models in critical scenarios such as medical triage and implemented Retrieval-Augmented Generation (RAG) and experimented with in-context learning to improve robustness and reliability of models.

Prior to starting my PhD, I did a Research Internship under Bloomberg Distinguished Professor Alan Yuille at John Hopkins University. During this time, I worked on robustness in vision and language models. I have also done a Research Assistantship at University of British Columbia under Professor Apurva Narayan. During this time, I worked on adversarial training in computer vision.

My academic journey began with a Bachelor's degree in Mathematics and Computing from IIT Delhi, India. After completing my bachelors, I have gained 2 and a half years of industry experience in Applied AI at HyperVerge, an innovative Fintech and Geospatial startup in India. During this time I worked on various projects on face recognition, face deduplication, OCR, forgery detection and object detection.

I am really interested with what AI can do in real-world settings. Currently I am very excited by NotebookLM podcasts by Google. I like giving my papers and let it generate a small podcast on it. Some on my publications have a podcasts linked with them now (credits to Google for that).

My research interests includes Responsible AI, evaluation, fairness and robustness.

When I am not doing research, I like to read books (currently reading lots of psychology and philosophy books) and travel.