I am an engineer and researcher with broad interests in Natural Language Processing, focusing on making LLMs safer and more reliable. My research spans AI Safety, Alignment, Reasoning, and Text Watermarking. Currently, I work as a Senior ML Engineer at Bloomberg where I focus on AI Safety and Hallucination Detection. I am also a part-time Ph.D. student at NJIT working with Prof Hai Phan. I am based in White Plains, just north of NYC .
Previously, I was an Applied Scientist at Amazon (2017-2023) where I built ML models for various NLP tasks and also had the opportunity to contribute to Amazon’s first LLM (AlexaTM-20B).
I received my Masters from Georgia Tech advised by Professor Le Song, focusing on Knowledge-Base Question Answering where we explored techniques such as Memory Networks and Graph-Embeddings. I also received the Donald V Jackson fellowship for outstanding Masters students.
Prior to grad school, I worked as an engineer at Bloomreach where I implemented scalable data pipelines in Map-Reduce. I obtained my Bachelors in Computer Science from IIT Ropar advised by Professor Apurva Mudgal on the Robot Localization problem. During undergrad, I interned with the Google Summer of Code program and was later voted as a top-level committer for the Apache Hama open‑source project for my contributions.
I am originally from Northern India . I spent my early childhood close to forests on the India-Nepal border. Before starting my Computer Science journey, I wanted to be a physicist. Outside research interests, my hobbies are reading non-fiction, learning new songs on the piano (brute-force), and photography.
Inspired by John Schulman, I maintain a notebook of research ideas; regrettably, I lack the time to pursue them all and would welcome some collaborators. I am an empirical researcher by practice, but I aim to ensure my findings are grounded in theory rather than purely heuristic observations. Whether you’re interested in collaboration (as equals, mentee, or mentor) or just want to discuss the math behind machine learning or the research meta-process and living a good life, let’s talk!
You can read more about my recent research on Red-Teaming & Threat Modeling and Watermarking & Alignment Mitigation.
― Professional Journey ―
