Systems-Facing Track – September 2025

Successes and challenges using LLMs within a Systems-Facing context

Speaker: Brandon Biggs, HPC Systems Administrator, Idaho National Lab

Abstract: This presentation shares practical insights from implementing large language models across Idaho National Laboratory with strict data governance requirements. We’ll explore the trade-offs between on-premises and cloud-based LLM deployments, including performance considerations, cost implications, and security controls. Key topics include navigating data residency requirements, balancing user accessibility with compliance needs, and lessons learned from our hybrid approach. Attendees will gain actionable strategies for LLM adoption in organizations where data protection is paramount.