About Me

Hello! I’m Advait Deshmukh, a master’s student in Computer Science at the University of Colorado Boulder with a keen interest in Natural Language Processing. I aim to leverage interpretability tools to study failure modes of language technologies and to make these systems more robust by instilling linguistically and cognitively informed inductive biases in them.

I’m currently conducting research at the BLAST Lab at University of Colorado Boulder, where I’m collaborating on various projects with Dr. Maria Pacheco. [Oct ‘25 Update: I’ll be presenting our work on Ultra-Fine Entity Typing at StarSEM’25 preprint — Please reach out if you’ll be attending EMNLP and would like to chat about open PhD positions in your lab :) or just say hi]. I recently collaborated with Rohan on his work ‘A Structured Clustering Approach for Inducing Media Narratives’ (Recommended to ACL ‘26 in July ARR, preprint).

These days, I am exploring ways to improve generalizability of popular Event Causality Identification (ECI) approaches with I-Ta and reframing Entity Typing problem in context of longer narratives with Maria.*

I have previously been involved in research at BAIC Lab led by Dr. Tom Yeh where I attempted to enhance vision capabilities in large language models though knowledge distillation, contributing to the advancement of multimodal AI research.

For my research activities at CU I received the prestigious Lloyd Botway Fellowship.

I’ve been part of the instructional staff for “CSCI 1300: Intro to Computing” at CU since Fall 2023. After my work as a course manager for the class, I was a TA for the class in Fall 2024/Spring 2025. I was awarded the best TA award by the department for my work with 1300.

Previously, I worked as a Software Engineer in the Innovation Team at Societe Generale Global Solution Center in Bangalore, where I contributed to the development of cutting-edge technological solutions, including a Retrieval-Augmented Generation (RAG) based chat system. I completed my undergraduate degree in Computer Science from BITS Pilani, where I gained a strong foundation in computer science and AI.

In addition to my professional and academic pursuits, I am passionate about reading books, playing chess, and painting cat-sitting, which provide me with a creative outlet and a way to unwind. If you’re interested in connecting or discussing potential collaborations, please feel free to reach out to me at advait.deshmukh@colorado.edu

* Come back later to find out about super-secret projects ;)

Education

  • M.S. in Computer Science
    CU Boulder, 2023 - 2025
    Boulder, Colorado, USA

  • B.E. in Computer Science
    BITS Pilani, 2017 – 2021
    Pilani, Rajasthan, India