Howdy! I am a Ph.D. candidate developing scalable algorithms and machine learning methods for graph and hypergraph networks. My work combines mathematical guarantees with practical implementations on large datasets. I also build agentic applications that combine LLMs with deterministic tools for computation and verification. Previously, I worked as a software engineer at HSBC and a data engineering intern at Walmart Global Tech.
I’m seeking research, applied scientist, and software engineering opportunities starting in 2027. Get in touch →
Research
Graphs describe pairwise connections. Hypergraphs describe group interactions. We develop algorithms for both.
Finding densely connected groups
We design algorithms to find densely connected groups in large networks. These methods account for interactions that involve several nodes at once. They balance speed with mathematical guarantees on solution quality.
Partitioning networks with group interactions
We study how to divide a network when splitting a group carries a cost. We develop efficient methods for these problems and establish limits on how well they can be solved.
Fast, flexible graph clustering
We develop fast algorithms that group related nodes in large networks. The methods support different levels of clustering detail. We also build tools to assess the quality of the resulting groups.
Predicting reciprocal group interactions
We study what it means to return an interaction when groups are involved. We define and measure these patterns. Then we test whether they help predict future group interactions.
Selected Projects
Applications of graph algorithms, machine learning, and AI systems.
RepoCompass — Explore and understand Python code
RepoCompass helps developers explore Python repositories and answer questions about their code. It combines code relationships with keyword and semantic search to find supporting evidence. On Django, it indexed 12,500 symbols and 8,000 call relationships. Validation includes 360 automated tests and manually labeled queries.
ResponseGuard-OT — Check incident-response plans
ResponseGuard-OT checks industrial incident-response plans against operational constraints. It suggests revisions when those constraints are violated. Operators can compare plans in simulated scenarios and approve each recommended action.
Finding candidate gene regulators in Shigella
I use graph clustering to find genes with similar expression patterns and identify potential regulators. I check whether candidates remain consistent across different analysis settings. The results support further study by experimental collaborators.
Publications
Also listed on Google Scholar.
Under review & revision
Measuring and Predicting Reciprocity in Hypergraph Interaction NetworksUnder review
Manuscript under review, 2026. Defines reciprocity in group interactions and tests its value for prediction.
Improved Hardness and Approximations for Cardinality-Based Minimum s–t Cuts in HypergraphsJournal revision
Under final journal revision, 2025.Studies the limits of efficient algorithms for partitioning networks with group interactions.
Peer-reviewed publications
A Survey on LLMs for Story Generation
Findings of the Association for Computational Linguistics: EMNLP 2025, pp. 13954–13966. Suzhou, China. Extended journal version under revision, 2026.Surveys how large language models are used to generate stories.
The Densest SWAMP Problem: Subhypergraphs with Arbitrary Monotonic Partial Edge Rewards
ECML-PKDD 2025 (research track, full paper). Porto, Portugal. Extended journal version under review.Finds dense groups while allowing partial rewards for group interactions.
Faster Approximation Algorithms for Parameterized Graph Clustering and Edge Labeling
CIKM 2023 (research track, full paper). Birmingham, UK.Develops fast graph clustering algorithms with guarantees on solution quality.
Experience
Full history in the CV.
Graduate Research Assistant, Texas A&M University
Aug 2022 – PresentAlgorithms and machine learning for large-scale graph and hypergraph data · College Station, TX
Develop graph algorithms in Julia and Python and prediction models in PyTorch. Evaluate them on large networks and release reproducible benchmarks and open-source code.
Data Engineer Intern, Walmart Global Tech
Jun 2021 – Aug 2021Cloud migration capacity planning · Dallas, TX
Built forecasting pipelines to estimate weekly compute demand. The results supported capacity planning for a Teradata-to-BigQuery migration.
Software Engineer, HSBC Software Development
Jul 2017 – Aug 2018Core banking platform · Pune, India
Designed and launched Autosolv, a Django search and Q&A tool for recurring AS400 build errors. It reduced repeat build failures by 15% and helped the team resolve build errors faster.
Teaching & Service
Texas A&M University.
- Teaching assistant in Machine Learning, Discrete Structures, Programming Languages, and Database Systems, supporting 200+ students. Built autograders and ML assignments. Helped students debug their projects.
- Reviewer: AAAI 2025, ECML-PKDD 2025, ALENEX 2026, ACM TIST 2026.
- Outreach: volunteer at Texas A&M CS Day (2023, 2024), including a small-group activity teaching sorting algorithms.
News
Talks, travel, and milestones.
-
Nov 16–20, 2026 upcoming
Giving a tutorial on hypergraph algorithms at the SIAM Conference on Mathematics of Data Science (MDS26) in Salt Lake City, Utah.
-
Oct 7–8, 2026
Presented a poster on reciprocity in hypergraph interaction networks at TL;DR 2026 (Texas Colloquium on Distributed Learning) at Rice University, Houston.
-
Jun 29–Jul 2, 2026
Gave a lightning talk on densest subhypergraphs at the ICERM workshop on Applied and Computational Discrete Algorithms in AI and Data Science in Providence, Rhode Island.
Earlier news
-
Dec 10, 2025
Passed my preliminary exam and Ph.D. thesis proposal.
-
Nov 4–6, 2025
Attended the Networks for Science workshop in Berkeley, California.
-
Sep 15–19, 2025
Presented a full paper (research track) at ECML-PKDD 2025 in Porto, Portugal.
-
Aug 20, 2025
Our survey paper on LLMs for story generation was accepted at Findings of EMNLP 2025.
-
Feb 20–22, 2025
Attended the ICERM workshop on Fusing Theory and Practice of Graph Algorithms.
- Nov 9, 2024
Volunteered at Texas A&M CS Day.
- Jul 2024
Teaching assistant for CSCE 310 — Database Systems.
- Jun 2024
Teaching assistant for CSCE 421 — Machine Learning.
- Oct 21–25, 2023
Presented a full paper (research track) at CIKM 2023 in Birmingham, UK.Develops fast graph clustering algorithms with guarantees on solution quality.
- Apr 13, 2023
Presented a poster at the TAMIDS Network Science workshop.
- Apr 1, 2023
Volunteered at CS Day and ran a small-group activity teaching students sorting algorithms.
- Mar 22, 2023
Passed my Ph.D. qualifying exam.