Tanny Chavez is a Computational Research Scientist at the Advanced Light Source, Lawrence Berkeley National Laboratory, where she serves as Program Lead for AI & Analytics within Photon Science Computing. Her research focuses on machine learning and software systems for scientific data, including AI-ready data systems, representation and feature extraction methods, and integrating agentic AI workflows into scientific data analysis pipelines. She holds a Ph.D. in Electrical Engineering from the University of Arkansas, where her doctoral research focused on segmentation methods for terahertz imaging.
Presentation Title:
Advancing AI-enabled scientific workflows at user facilities
Presentation Abstract:
The growing scale, diversity, and complexity of data produced across instruments at scientific user facilities create new opportunities for AI-enabled analysis. At the Advanced Light Source, we are working toward integrated capabilities for incorporating AI-enabled analysis into experimental workflows, connecting data acquisition, computational infrastructure, metadata and provenance, and reproducible analysis pipelines. These efforts also include exploring agentic AI approaches that can support experimental planning and execution while operating within clearly defined instrument interfaces and constraints. Together, these capabilities aim to streamline the path from measurement to insight, reduce analysis bottlenecks, improve reproducibility, and support the reuse of AI methods across beamlines, techniques, and scientific domains. This work highlights practical considerations for building scalable AI infrastructure in multidisciplinary user-facility environments.