Shikhar Srivastava

I'm a Computer Science PhD student at the University of Rochester, advised by Christopher Kanan.

I graduated in Computer Science from MIT Manipal with the Institute's Academic Excellence Award.
I then led data science and the AI Roadmap for DHL Ecommerce's entire Smart-Trucking India business.
I also did some pre-doctoral work on continual learning for healthcare at MBZUAI & G42.

I'm doing a PhD to enable large models to learn continually: I'm primarily interested in memory, learning dynamics, and enabling non-negative knowledge transfer across shared task structures in LLMs.

I also work with Catherine Arnett at EleutherAI, and Dhireesha Kudithipudi, Itamar Lerner at UTSA.

Publications

LayerRoPE: Dynamic Depth-wise Magnitude & Angular Superposition
S. Srivastava, C. Kanan
COLM 2026, Actionable Interpretability
No Single Tokenizer Feature Reliably Predicts Downstream Language Model Performance
A. Lopardo, L. F. Salim, S. Srivastava, K. Jiang, M. Sharma, C. Arnett
EMNLP 2026, Findings
SHARP: Sleep-based Hierarchical Accelerated Replay for Long Range Non-Stationary Temporal Pattern Recognition [pdf]
J. Dey, S. Srivastava, I. Lerner, C. Kanan, D. Kudithipudi
CoLLAs 2026
The Unreasonable Benchmark [pdf]
F. Perez-Cruz, B. Hitaj, G. M. Dimitri, O. Deb, A. M'Charrak, S. Srivastava, C. Kanan, et al.
DMLR 2026
Improving Multimodal Large Language Models Using Continual Learning [pdf]
S. Srivastava, M. Y. Harun, R. Shrestha, C. Kanan
CoLLAs 2025 · NeurIPS 2024, Scalable Continual Learning for Lifelong Foundation Models
Revisiting Multi-Modal LLM Evaluation [pdf]
J. Lu*, S. Srivastava*, J. Chen, R. Shrestha, M. Acharya, K. Kafle, C. Kanan
CVPR 2025, Benchmarking & Expanding AI Multimodal Approaches · ICLR 2025, Navigating and Addressing Data Problems for Foundation Models
Continual Learning for Chest X-Ray Classification in Low-Resource Clinical Settings [pdf]
S. Srivastava, M. Yaqub, K. Nandakumar, Z. Ge, D. Mahapatra
MICCAI 2021, FAIR · Best Paper Award

* Equal contribution.

Personal

I'm fascinated by memory. By the phenomena arising from our brains developing collectively through the phylogenetic tree. By old stories and old ideas.

I believe strongly in working towards the benefit of all.

***