Event Details

Talk Abstract:

LoomAI is a new open-source, browser-based interface that transforms how researchers interact with the FABRIC testbed. As the newly recommended way to use FABRIC, LoomAI replaces command-line and Jupyter notebook workflows with an intuitive, visual, point-and-click experience for designing, deploying, and managing experiments.

A central focus of this session is LoomAI’s integration with the SPHERE testbed: from a single interface, researchers can design and orchestrate experiments across both FABRIC’s nationwide and international infrastructure and SPHERE’s environment for reproducible cybersecurity and privacy research. We will showcase the drag-and-drop topology editor for visually building network experiments, then walk through how the same workflow extends across facilities.

Attendees will also see LoomAI’s built-in tools, including SSH terminals, a file manager, and a geographic resource map, and its integrated AI: an assistant for experiment design, troubleshooting, and analysis, plus AI coding tools for deeper development. We will introduce “Weaves,” reusable automated experiment packages, and the Artifact Marketplace for sharing community-driven templates. All AI features are powered by models hosted on FABRIC and National Research Platform GPUs, ensuring a secure, research-focused environment with no reliance on commercial infrastructure.

Whether you are new to FABRIC or modernizing your workflow, this session shows how LoomAI can accelerate your research from start to finish.

Speaker Bio

Komal Thareja is a Senior Research Software Engineer at RENCI, University of North Carolina at Chapel Hill, where she leads the development of FABRIC's core platform capabilities — from advanced networking services and heterogeneous accelerators to distributed storage and experiment orchestration. Her current research explores the convergence of distributed intelligence, multi-agent systems, programmable networks, and AI-driven scientific workflows, with a focus on enabling large-scale, reproducible experimentation across edge, network, and cloud environments. She brings years of industry experience architecting large-scale systems in the communications sector, grounding her research in production-grade engineering discipline.