Tarun Kumar

Senior Research Scientist

HPE Labs

kumartaruncse [at] gmail [dot] com

I am a Senior Research Scientist at Hewlett Packard Enterprise (HPE) Labs, where I work on Generative AI, Large Language Models (LLMs), agentic AI systems, and Natural Language Processing. My work focuses on translating novel algorithms into enterprise-scale AI systems — spanning retrieval-augmented generation (RAG), LLM agent guardrails, foundation model virtualization, and model selection pipelines.

I earned my Ph.D. in Computer Science and Engineering from IIT Madras as an Intel India Ph.D. Fellow, advised by Balaraman Ravindran and Manikandan Narayanan. My thesis, Algorithms for Rich Graph Structures, received the Best Ph.D. Thesis award in Data Science at IIT Madras.

I have authored papers at venues including ICML, ACL, and AAAI, and have collaborated across teams. My earlier research contributed to hypergraph and multilayer network learning, with applications in network biology.

Feel free to explore my publications, research, and CV, and reach out if you would like to collaborate.

news

Jul 01, 2026	I am co-organizing the AgenticOS workshop at NeurIPS 2026, focused on operating-system abstractions for agentic AI.
May 01, 2026	Our paper “It is Time to Virtualize Foundation Models with a Self-evolving Operating System Layer” has been accepted at ICML 2026. 🎉
Feb 01, 2026	Honored to receive the Distinguished Alumni Award from Shiv Nadar University (2026).
Jan 25, 2026	Co-organizing the GCLR workshop series (through AAAI 2026)
Dec 10, 2025	Our work on InfrastructureSentinel: Policy-Enforced Guardrails for Secure MCP-driven Infrastructure Agents was accepted at AAAI 2026, with a companion paper at the AAAI 2026 workshop on agentic AI benchmarks.

selected publications

ICML

It is Time to Virtualize Foundation Models with a Self-evolving Operating System Layer

Suparna Bhattacharya, Tarun Kumar, Cong Xu, and 7 more authors

In International Conference on Machine Learning (ICML), 2026

Bib

@inproceedings{bhattacharya2026virtualize,
  title = {It is Time to Virtualize Foundation Models with a Self-evolving Operating System Layer},
  author = {Bhattacharya, Suparna and Kumar, Tarun and Xu, Cong and Mopur, Satish Kumar and Li, Jiahao and Mishra, Ashish and Tripathy, Aalap and Koomthanam, Annmary Justine and Foltin, Martin and Foster, Ian},
  booktitle = {International Conference on Machine Learning (ICML)},
  year = {2026},
}

ACL

CEBC: Conformal Evidence-Bounded Control for Low-Hallucination Vision–Language Generation

Ashish Mishra, Tarun Kumar, Arpit Shah, and 2 more authors

In Annual Meeting of the Association for Computational Linguistics (ACL), 2026

Bib

@inproceedings{mishra2026cebc,
  title = {CEBC: Conformal Evidence-Bounded Control for Low-Hallucination Vision--Language Generation},
  author = {Mishra, Ashish and Kumar, Tarun and Shah, Arpit and Bhattacharya, Suparna and Foltin, Martin},
  booktitle = {Annual Meeting of the Association for Computational Linguistics (ACL)},
  year = {2026},
}

AAAI

InfrastructureSentinel: Policy-Enforced Guardrails for Secure MCP-driven Infrastructure Agents

Tarun Kumar, Aalap Tripathy, Gayathri Saranathan, and 5 more authors

In Proceedings of the AAAI Conference on Artificial Intelligence, 2026

Bib

@inproceedings{kumar2026infrastructuresentinel,
  title = {InfrastructureSentinel: Policy-Enforced Guardrails for Secure MCP-driven Infrastructure Agents},
  author = {Kumar, Tarun and Tripathy, Aalap and Saranathan, Gayathri and Bhattacharya, Suparna and Foltin, Martin and Hinchley, Scott and Bahls, Donald M and Brookshire, David},
  booktitle = {Proceedings of the AAAI Conference on Artificial Intelligence},
  volume = {40},
  number = {47},
  pages = {40295--40301},
  year = {2026},
}

ACL

SubLIME: Subset Selection via Rank Correlation Prediction for Data-Efficient LLM Evaluation

Gayathri Saranathan, Cong Xu, Mahammad Parwez Alam, and 4 more authors

In Annual Meeting of the Association for Computational Linguistics (ACL), 2025

Best Theme Paper Bib

Best Theme Paper Award; Oral presentation (top 1% of accepted papers).

@inproceedings{saranathan2025sublime,
  title = {SubLIME: Subset Selection via Rank Correlation Prediction for Data-Efficient LLM Evaluation},
  author = {Saranathan, Gayathri and Xu, Cong and Alam, Mahammad Parwez and Kumar, Tarun and Foltin, Martin and Wong, Soon Yee and Bhattacharya, Suparna},
  booktitle = {Annual Meeting of the Association for Computational Linguistics (ACL)},
  year = {2025},
}

PLOS Comp. Bio.

MultiCens: Multilayer Network Centrality Measures to Uncover Molecular Mediators of Tissue-Tissue Communication

Tarun Kumar, Ramanathan Sethuraman, Sanga Mitra, and 2 more authors

PLOS Computational Biology, 2023

Bib

@article{kumar2023multicens,
  title = {MultiCens: Multilayer Network Centrality Measures to Uncover Molecular Mediators of Tissue-Tissue Communication},
  author = {Kumar, Tarun and Sethuraman, Ramanathan and Mitra, Sanga and Ravindran, Balaraman and Narayanan, Manikandan},
  journal = {PLOS Computational Biology},
  volume = {19},
  number = {4},
  pages = {e1011022},
  year = {2023},
  publisher = {Public Library of Science},
}