IRIS-HEP Fellow: Arvind Tawker

Fellowship dates: Jul – Sep, 2025
Home Institution: University of California, Davis
Project: Development of Modular ML Pipeline Templates for High Energy Physics Applications
High Energy Physics (HEP) research relies increasingly on machine learning (ML) to manage the vast, complex datasets produced by modern experiments. Techniques such as deep neural networks, autoencoders, generative models, and graph-based networks have become essential for tasks ranging from signal/background classification and energy regression to anomaly detection and fast simulation. Yet despite many successes, ML pipelines in HEP remain largely ad hoc—leading to duplicated effort and inconsistent practices—just as the High-Luminosity upgrades promise even larger, more detailed datasets. To address this, we will develop a comprehensive GitHub repository of modular ML pipeline templates, where each branch hosts a self-contained workflow tailored to a common HEP use case. By embedding best practices in modular code design, reproducibility, experiment tracking, CI/CD (with pre-commit hooks), and ROOT data handling via Uproot, these templates will dramatically reduce onboarding time and promote standardized, production-ready ML workflows across collaborations.
More information: My project proposal
Mentors:
-
Dr. Liv Vage (Princeton University)
- - "", Arvind Tawker,
Current Status
Contact me: