Research

My work sits at the intersection of streaming graph algorithms, distributed systems, and GPU-accelerated computing. The unifying theme: using graphs as a first-principles framework to make large-scale computation fast, correct, and practical.

SAGA — State-Aware Streaming Graph Analytics #

Large graphs change continuously — edges are inserted and deleted at high throughput. Most systems recompute analytics from scratch on each update, which doesn’t scale. SAGA is a distributed framework that maintains analytical state incrementally over dynamic graphs, propagating only the changes that matter across partitions.

Accepted · HPDC 2026

StreamGC — Real-Time Vertex Coloring on Parallel Systems #

Vertex coloring assigns labels to nodes such that no two adjacent nodes share a label. In the streaming setting, edges arrive in batches and the coloring must be maintained without global synchronization. StreamGC handles this on parallel systems with low update latency.

Past Work #

Semantic PageRank (M.Tech Thesis, IIT Jodhpur · 2022–2024)
Extended classical PageRank with knowledge graph signals to improve ranking quality on academic corpora. Applied distributed graph processing on Apache Spark GraphX.