Shwetha Somasundaram

Research Associate II

Bangalore

I am a Research Associate at Adobe Research, India. I am broadly interested in machine learning, natural language processing and related fields. My recent projects include developing a training-free speculative decoding technique for contextual tasks,  retrieval and attribution methodologies for document question answering,  and foundation models for design generation/editing.

Prior to joining Adobe, I graduated from SSN College of Engineering in 2022 with a Bachelor’s degree in Electronics and Communication Engineering. During my undergraduate study, I interned at Adobe Research and developed a domain-specific Named Entity Recognition (NER) system for low-resource settings. I also interned at the Medical Imaging Lab, Indian Institute of Science, Bangalore where I worked on estimating Tracer Kinetic parameters from under-sampled DCE-MRI data.

My personal website can be found here.

Publications

PLD+: Accelerating LLM Inference by Leveraging Language Model Artifacts

Somasundaram, Shwetha., Phukan, Anirudh., Saxena, Apoorv. (Apr. 29, 2025)

Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics (NAACL)

Peering into the Mind of Language Models: An Approach for Attribution in Contextual Question Answering

Phukan, Anirudh., Somasundaram, Shwetha., Saxena, Apoorv., Goswami, Koustava., Srinivasan, Balaji. (Aug. 12, 2024)

Findings of the Association for Computational Linguistics (ACL 2024)

Presentations by the Humans and For the Humans: Harnessing LLMs for Generating Persona-Aware Slides from Documents

Mondal, Ishani., Somasundaram, Shwetha., Natarajan, Anandhavelu., Garimella, Aparna., Bandyopadhyay, Sambaran., Boyd-Graber, Jordan. (Mar. 17, 2024)

The 18th Conference of the European Chapter of the Association for Computational Linguistics (EACL) 2024

Drilling Down into the Discourse Structure with LLMs for Long Document Question Answering

Nair, Inderjeet., *, *., Somasundaram, Shwetha., *, *., Saxena, Apoorv., Goswami, Koustava. (Dec. 6, 2023)

Conference on Empirical Methods in Natural Language Processing (EMNLP)