AGI Research Scientist

Swastik Roy

Research notes on AI systems, evaluation, and alignment

Swastik Roy works on AI systems, evaluation, alignment, post-training, and reliable measurement.

github linkedin

Research Interests

LLM evaluation and model behavior
Model specifications, AI constitutions, and alignment
Post-training, reward models, rubrics, judges, and RL environments
Reliable benchmark and evaluation methodology

Recent Publications

JudgeBoard: Benchmarking and Enhancing Small Language Models for Reasoning Evaluation

Zhenyu Bi, Gaurav Srivastava, Yang Li, Swastik Roy, Meng Lu, Morteza Ziyadi, Xuan Wang · Proceedings of the AAAI Conference on Artificial Intelligence · 2026

Publication source

BeyondBench: Contamination-Resistant Evaluation of Reasoning in Language Models

Gaurav Srivastava, Aafiya Hussain, Zhenyu Bi, Swastik Roy, Priya Pitre, Meng Lu, Morteza Ziyadi, Xuan Wang · arXiv (Cornell University) · 2025

Publication source

JudgeBoard: Benchmarking and Enhancing Small Language Models for Reasoning Evaluation

Zhenyu Bi, Gaurav Srivastava, Yang Li, Meng Lu, Swastik Roy, Morteza Ziyadi, Xuan Wang · ArXiv.org · 2025

Publication source

OPTAGENT: Optimizing Multi-Agent LLM Interactions Through Verbal Reinforcement Learning for Enhanced Reasoning

Zhenyu Bi, Meng Lu, Yang Li, Swastik Roy, Weijie Guan, Morteza Ziyadi, Xuan Wang · ArXiv.org · 2025

Publication source

OPTAGENT: Optimizing Multi-Agent LLM Interactions Through Verbal Reinforcement Learning for Enhanced Reasoning

Zhenyu Bi, Meng Lu, Yang Li, Swastik Roy, Weijie Guan, Morteza Ziyadi, Xuan Wang · Proceedings of the 14th International Joint Conference on Natural Language Processing and the 4th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics · 2025

Publication source

SPARK: Stepwise Process-Aware Rewards for Reference-Free Reinforcement Learning

Salman Rahman, Sruthi Gorantla, Arpit Gupta, Swastik Roy, Nanyun Peng, Yang Liu · ArXiv.org · 2025

Publication source