Fifth Annual Workshop on Cyber Security in High Performance Computing (S-HPC'26)


Program
McCormick Place Chicago, Illinois, USA
Room W183b

TimeEvent
14:00 - 14:03Welcome:
Presenter: Joseph Manzano (Pacific Northwest National Laboratory)
14:03 - 14:22Technical Paper: A Threat Model for Sensitive-Data Workloads on Multi-Tenant AI Supercomputers
Sadaf R. Alam, Martin O'Reilly, Matt Craddock, Jim Madge, Matt Penn, George Svarovsky, Jake Watson, Babatunde Oyewo
14:22 - 14:41Technical Paper: Privacy Foundations for Multi-Institutional Scientific Artificial Intelligence
Olivera Kotevskai, Sumit Jha, Aurélien Bellet, Rui Hu, Nathaniel Bastian, Rafael Ferreira da Silva, Ravi Madduri, Kibaek Kim
14:41 - 15:00Technical Paper: A First Look at Conventional Attacks on Scientific Applications
Michael Polinski, Peter Dinda
15:00 - 15:30Break
15:30 - 16:03Distinguished Speaker: Rethinking Security for AI: Enforcing Boundaries for AI Agents
Tim Klawa (NVIDIA)
16:03 - 16:22Technical Paper: TMD - A Label-Free, Misalignment-Robust Distance for HPC Attack Detection
Venkat Sai Suman Lamba Karanam, Zahmeeth Sakkaff, Shubham Sundriyal
16:22 - 16:41Technical Paper: Quantifying Workload Exposure in HPC Power Telemetry at Operational Sampling Rates
Kwesi Danso, Samuel Abiola, Susan Mengel
16:41 - 17:00Technical Paper: Towards Distributed-Memory All-Pairs Reachability over Homomorphically Encrypted Graphs
Charis Hului, Anrin Chakraborti, Sidharth Kumar
17:00 - 17:30Panel
Moderator: Yang Guo (NIST)

HPC Security in the AI Era: Standards, Compliance, and Governance
Panelists: Ryan Adamson (Oak Ridge National Laboratory); Sadaf Alam (University of Bristol); Rickey Gregg (DoW/HPCMP)
17:30 - 17:31Closing Remarks
Joseph Manzano, Pacific Northwest National Laboratory

Distinguished Speaker: Tim Klawa, NVIDIA

 

Title: Rethinking Security for AI: Enforcing Boundaries for AI Agents

 

Abstract

As organizations adopt AI agents that execute code, access data, and coordinate across shared HPC infrastructure, every allowed action becomes a security decision. This talk explores how NVIDIA OpenShell brings isolation, deterministic policy enforcement, and controlled network access, unlocking actual enforceable agent runtime boundaries. Safe adoption requires extending zero trust to agent behavior, making continuous monitoring and governance foundational to how organizations grant, oversee, and govern agent autonomy.

Bio:

Tim Klawa is a Public Sector Field Engineering Leader at NVIDIA, where he helps federal customers and partners design and implement accelerated AI solutions with a focus on cybersecurity. Prior to joining NVIDIA, Tim served as Public Sector CTO at Groq. Throughout his career, he has focused on architecting and optimizing AI systems that transform complex data into actionable intelligence. His experience spans building nation-scale data management platforms, optimizing AI-driven analytics, to orchestrating agent swarms in a safe and governed way.

Designed with Mobirise web theme