Ph.D. student, Computer Science, Stony Brook University. Expected May 2027. Open to backend and infrastructure software engineering roles in the US.
About
I build high-performance C++ systems for streaming and out-of-core graphs.
At Stony Brook I led the implementation of GraphZeppelin
and CubeSketch: a streaming engine that applies millions of edge updates per second
and computes connectivity on dense, dynamic graphs that do not fit in RAM.
I also supervised undergraduates on that codebase.
I am looking for backend, storage, and infrastructure roles that use algorithms and systems software.
Built GraphZeppelin, a C++ streaming graph engine (2–5M edge updates/sec) for connectivity on graphs larger than RAM. Code
Implemented CubeSketch, an ℓ0 sketch used in that engine: 4× less space and about 1000× faster updates than prior samplers for this workload.
Added out-of-core I/O (node-based buffering, gutter trees) so the same system can run on SSD for much larger graphs, with about 29% throughput loss.
Parallelized ingestion with OpenMP (46 threads, 25× throughput) and built Linux cgroup memory-pressure tests for cache-oblivious sorting and matrix multiplication.
Teaching Assistant
Stony Brook University | 2017-2023
TA for 9 CS courses (30-105 students), including Algorithms, Software Engineering, Operating Systems, and graduate NLP.
Built Linux sandboxing for autograding (cgroups, ulimit).
Implemented neural models for semantic role labeling in Saul, a Java/Scala ML framework that treats inputs as graphs.
Publications
David Tench, Evan West, Victor Zhang, Michael A. Bender, Abiyaz Chowdhury, J. Ahmed Dellas, Martin Farach-Colton, Tyler Seip, and Kenny Zhang.
GraphZeppelin: Storage-Friendly Sketching for Connected Components on Dynamic Graph Streams.SIGMOD 2022.
David Tench, Evan West, Victor Zhang, Michael A. Bender, Abiyaz Chowdhury, Daniel Delayo, Martin Farach-Colton, Tyler Seip, and Kenny Zhang.
GraphZeppelin: How to Find Connected Components (Even When Graphs Are Dense, Dynamic, and Massive).ACM TODS 2024.
Arghya Bhattacharya, Abiyaz Chowdhury, Helen Xu, Rathish Das, Rezaul A. Chowdhury, Rob Johnson, Rishab Nithyanand, and Michael A. Bender.
When Are Cache-Oblivious Algorithms Cache Adaptive? A Case Study of Matrix Multiplication and Sorting.ESA 2022.