- ICML 2025 position paper on data leakage in graph matching benchmarks (with proposals for mitigation)
- Serving as a reviewer for NeurIPS 2025, ICLR 2026
Research Experience
- Visiting Research Assistant with Vikas Garg at QuML Lab @ Aalto University
- Co-organizing Workshop on Differentiable Learning of Combinatorial Algorithms (DiffCoALG@NeurIPS 2025)
- Presenting a tutorial on Retrieval of Graph Structured Objects: Theory and Applications at CIKM 2025
Education
PhD student at CSE@IIT Bombay, jointly advised by Abir De and Soumen Chakrabarti.
Background
Research interests include designing machine learning methods for predictive challenges on graphs, focusing on neural graph retrieval. This involves developing deep learning models to solve graph combinatorial problems, scalable retrieval methods with trainable hashing, and ensuring interpretability through explicit alignment-driven justifications. These techniques are applicable to multimodal retrieval involving graphs from diverse domains, such as knowledge graphs, scene graphs, social networks, and molecular graphs.