Bio
Emily Herron is a Postdoctoral Research Associate in the Analytics & AI Methods at Scale (AAIMS) Group in the National Center for Computational Science at Oak Ridge National Laboratory. Her research sits at the intersection of large-scale AI and scientific discovery, spanning neural architecture search, foundation model training and scaling, and agentic AI systems, with a growing focus on the safety, security, and trustworthiness of autonomous AI in high-stakes scientific environments. Her current work centers on agentic AI and its trustworthiness, including VISTA, an agentic retrieval system for molten-salt thermophysical-property data built under the DOE Genesis Mission; VISTAGuard, a security and safety-alignment framework with capability-aware safety gates and a companion red-teaming benchmark (SciAgentBench); CHUNKS, a multi-agent spectroscopic-analysis pipeline; and SciTrust, a framework for evaluating truthfulness, adversarial robustness, safety, and ethics in scientific AI. She also developed HARMONY, a NAS framework for hybrid Transformer-Mamba-MoE models scaling to 16,384 GPUs on Frontier. in Data Science & Engineering from the Bredesen Center at the University of Tennessee and a B.S. in Computational Science from Mercer University.