I’m Kushal, an AI researcher working on Deep Learning on Graphs. I am also interested in handling data in non-Euclidean spaces.

I’m currently a final year CS PhD student at the Indian Statistical Institute, supervised by Prof. Swagatam Das. During my Phd tenure, I have worked on various problems in long-range interactions like Oversmoothing, Oversqaushing, and Graph Heterophily. I have also explored the domain of Graph Transformers and various Positional Encodings.

News!

1. [August 2025] Nominated as Reviewer in the 4th Learning on Graph Conference (LoG) for the Main Track.
2. [August 2025] Our joint work titled FairSplit: Mitigating Bias in Graph Neural Networks through Sensitivity-based Edge Partitioning accepted in the ACM CIKM 2025 in the short research paper track.
3. [July 2025] Nominated as Reviewer for NeurIPS, 2025 for Position Track.
4. [April 2025] Our joint work titled Can Graph Neural Networks Tackle Heterophily? Yes, With a Label-Guided Graph Rewiring Approach! accepted in IEEE Transactions on Neural Networks and Learning Systems (IEEE TNNLS).

Research Interests

1. Message-passing Graph Neural Networks
2. Graph Transformers and Positional Encodings
3. Oversmoothing and Oversquashing
4. Graph Fairness