Saransh Sharma

Research Associate

Bangalore

I’m Saransh Sharma, a 2025 Dual Degree (B.Tech + M.Tech) graduate in Computer Science from IIT Kharagpur. My research focuses on designing intelligent document experiences, including developing new document interactions and functionalities, extracting multimodal insights, and generating multimodal artifacts that help users better understand, explore, and interact with document content. My thesis focused on leveraging large language models for intent detection in both text and multimodal settings. I have also worked on agentic frameworks for extractive summarization and enhancing multilingual understanding of language models.

Publications

An Answer is just the Start: Related Insight Generation for Open-Ended Document-Grounded QA

Sharma, Saransh., Garimella, Aparna., Mukherjee, Koyel., Sharma, Sharma. (Jul. 2, 2026)

Association for Computational Linguistics Findings (ACL 2026)

FiRST: Finetuning Router-Selective Transformers for Input-Adaptive Latency Reduction

Jain, Akriti., Sharma, Saransh., Mukherjee, Koyel., Pal, Soumyabrata. (Nov. 4, 2025)

Conference on Empirical Methods in Natural Language Processing Findings (EMNLP 2025)

Text Takes Over: A Study of Modality Bias in Multimodal Intent Detection

Sharma, Saransh. (Nov. 4, 2025)

Conference on Empirical Methods in Natural Language Processing (EMNLP 2025)