ICLR 2026, Rio de Janeiro, Brazil
I am a Ph.D. student in Computer Science and Engineering at The Ohio State University, where I am fortunate to be advised by Prof. Yu Su. I previously earned an M.Sc. in Computer Science (thesis track) from Université de Montréal, affiliated with Mila. Before beginning my doctoral studies, I worked as a Software Engineer at IBM India Research Lab in Bangalore. I hold a Bachelor of Technology (Honours) from IIT Kharagpur, where I was awarded the prestigious Institute Silver Medal. I have also had the privilege of spending three summers as a research intern at Microsoft Research, Google Brain, and Bosch AI Research.
My research interests lie in LLM agents, multimodal foundation models, and KB reasoning, with a central focus on the role of structure in modern multimodal machine learning. Broadly, I study how leveraging structure, either inherent (e.g., knowledge graphs) or deliberately synthesized (e.g., web interaction trajectories), can make multimodal systems more robust, generalizable, and interpretable. My work integrates visual signals, language, and structured knowledge to enhance representation and reasoning (ACL'21, CIKM'23, CIKM'24) as well as interactive web environments (ACL (Findings)'25). I also work on computer-use agents and LLM interpretability, including the use of sparse autoencoders to analyze salient latent structure in dense image representations (ICLR'26). My earlier work spans complex sequential question answering over knowledge graphs (AAAI'18) and multi-task sequence labeling with recurrent neural networks (Interspeech'17).
Reviving the Context: Camera Trap Species Classification as Link Prediction on Multimodal Knowledge Graphs
[pdf]
[code]
CIKM 24 (Oral)
[poster] Honourable Mention, OSU CSE Graduate Student Research Poster Exhibition 2024
[short version] CV4Animals workshop, CVPR 24 (Oral)
Fine-Tuning is Fine, if Calibrated
[pdf] NeurIPS 2024
Holistic Transfer: Towards Non-Disruptive Fine-Tuning with Partial Target Data
[pdf] NeurIPS 2023
Knowledge Base Question Answering: A Semantic Parsing Perspective
[pdf] AKBC 2022
Learning a Probabilistic Boolean Network Model from Biological Pathways and Time-series Expression Data
[paper] EMBC 16
SalsaBot: Towards a Robust and Generalizable Embodied Agent
[short version]
Embodied AI Workshop at CVPR 2023
[long version]
Alexa Prize SimBot Challenge Proceedings 2023
Learning Sparse Mixture of Experts for Visual Question Answering
[pdf] Visual Question Answering and Dialog Workshop, CVPR 2019
Tooling framework for instantiating natural language querying system
[pdf] VLDB Endowment 2018
ICLR 2026, Rio de Janeiro, Brazil
OSU NLP group
ACL 2025, Vienna, Austria
Research Trends forum
CV4Animals Workshop, CVPR 2024, Seattle, WA
CIKM 2023, Birmingham, UK
Guest lecture, OSU, CSE 6521