Hey! I'm Shubham Gandhi.
I am an AI/NLP graduate student at the Language Technology Institute at Carnegie Mellon University (CMU). Currently, I am working with Prof. Carolyn Rose, Yiqing Xie, and Atharva Naik on large language models (LLMs) for automated code onboarding, where we are establishing a benchmark for question-answering with LLMs for repository-level onboarding.
Prior to CMU, I was working as a Research Fellow at TCS Research with Dr. Manasi Patwardhan on developing LLM agents for the automated codification of research problems and experiments, focusing on the entire end-to-end research pipeline. Here, I have also worked on code translation for low-resource programming languages using LLMs, self-refinement, and high-resource languages as a pivot.
In 2023, I completed a B.E in Computer Science with a Minor in Data Science from Birla Institute of Technology and Science, Pilani - Goa Campus. Here, I began my research journey exploring the depths of AI, from the complexities of deep learning, adversarial robustness, and large language models to forefront challenges in AI4Code. I aim to not only further my understanding of niche concepts such as LLMs but also to strategically wield this diverse expertise. My goal is to transform these experiences into tangible contributions that resonate within the cutting-edge arenas of AI research including, but not limited to, nuanced applications of LLMs in the field of code.
In the past, I have had the pleasure of working in both industrial and academic research settings. At Microsoft Research, I worked on automated program repair using static analysis with Dr. Sriram Rajamani, Dr. Nagarajan Natarajan and Suresh Parthasarathy among others. I have also worked with Dr. Gautam Shroff on applications of LLMs for time-series tabular data. Apart from this, I have also had experience in other domains, including computer vision, adversarial ML and brain-inspired DL as part of various academic research labs.
I'm always open to challenging research opportunities in AI that push me out of my comfort zone, which would benefit my research career.