Portrait of Swaroop Nath

Swaroop Nath

Member of Technical Staff @ Reflection AI

Experience

Member of Technical Staff

Reflection AI

May 2026 – Present

Researcher

Google DeepMind

July 2024 – May 2026

AI/ML Research Intern

LinkedIn

May 2023 – July 2023

Education

MS by Research, CSE

Indian Institute of Technology, Bombay

Aug 2021 – June 2024

About

Hi! I am a Member of Technical Staff at Reflection AI, working on enabling long-horizon reasoning and agentic capabilities.

Before this, I spent two years at Google DeepMind, in the Machine Learning and Optimization team led by Dr. Prateek Jain, where I worked on Efficiency — both in Pre-training and Post-training. My most significant work was on making Gemini's reasoning process efficient (more details in CV).

I graduated as an MS by Research scholar from the CFILT Lab, IIT Bombay. I was fortunate enough to be advised by Dr. Pushpak Bhattacharyya (God rest his soul) and Dr. Harshad Khadilkar. My Thesis was recognized with the Winifred B. Fernandes Research Excellence Award. My works have been featured in A* NLP conferences — EMNLP and ACL.

Schedule a Conversation

I am happy to set aside 15 minutes for a conversation — on anything from feedback on an early idea to advice on your career, or a potential collaboration. If we have not met before, I would appreciate a brief introductory email ahead of the meeting.

Schedule a 15-minute conversation

News

Selected Publications (* = equal contribution)

Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities

Google DeepMind (large-scale collaboration)

Technical Report · arXiv

Leveraging Domain Knowledge for Efficient Reward Modelling in RLHF

Swaroop Nath*, Tejpalsingh Siledar*, Pushpak Bhattacharyya, et al.

arXiv Preprint · arXiv · Code · Slides

One Prompt To Rule Them All: LLMs for Opinion Summary Evaluation

Swaroop Nath*, Tejpalsingh Siledar*, Pushpak Bhattacharyya, et al.

ACL 2024 · Paper · Code

Transformers are Expressive, But Are They Expressive Enough for Regression?

Swaroop Nath, Harshad Khadilkar, and Pushpak Bhattacharyya

arXiv Preprint · arXiv · Code

Reinforcement Replaces Supervision: Query focused Summarization using Deep Reinforcement Learning

Swaroop Nath, Pushpak Bhattacharyya, and Harshad Khadilkar

EMNLP 2023 · Paper · Code · Slides