2026 Midwest RCD Consortium Annual Network Meeting

This session provides a beginner-friendly introduction to quantum computing, highlighting how it differs from classical computers. Using intuitive, real-world examples, participants will learn core concepts such as superposition and entanglement, without requiring any math or physics background. The session explains why quantum computing matters, the types of problems it is suited for, and current limitations. Attendees will explore practical ways to begin experimenting with quantum computing today using free, publicly available tools and simulators. By the end of the session, participants will gain a conceptual understanding of quantum computing, hands-on experience with accessible tools, and confidence to explore this emerging technology further. This session is highly relevant to RCD practitioners, as it demonstrates how quantum computing concepts can inspire new approaches to research, creativity, and problem-solving in diverse disciplines. Key Takeaways: 

- Understand the difference between classical and quantum computing.
- Grasp fundamental quantum concepts such as superposition and entanglement.
- Recognize the types of problems quantum computing can address.
- Explore free tools and resources for hands-on experimentation with quantum computing.

• 9:45 am | Talk 1: Hello Computer: HPC in the Agentic Era (15 mins)
Speaker: Geoffrey Lentner, Purdue University

Focus: Geoffrey will share a practitioner’s perspective on how HPC centers are adapting to the rapid rise of agentic AI, systems that can reason, plan, and execute multi-step tasks. Drawing on work at Purdue’s Rosen Center for Advanced Computing, he will discuss deploying purpose-built MCP servers (HPC + Globus), developing system-wide configurations, and guiding researchers who want agentic tools integrated into their workflows. He’ll highlight early productivity gains, lessons learned (including writing a conference paper with agentic workflows), and an emerging framework for balancing enablement with responsible oversight. Key takeaways: 
• How agentic AI differs from earlier AI approaches
• Practical use of agentic tooling in HPC environments

• 10:00 am | Talk 2: Garden: Democratizing Access to AI/ML Models for Scientific Research (15 mins)
Speaker: Owen Price-Skelly, The University of Chicago

Focus: Owen will introduce Garden, an open platform designed to close the gap between AI/ML model publication and practical reuse. Garden enables researchers to publish containerized models, with metadata, benchmarks, training data, and documentation, and lets users run them in just a few lines of Python without managing dependencies, GPUs, or infrastructure. With integrations for cloud execution (Modal) and HPC (Globus Compute), Garden supports reproducible, community-curated model collections that evolve into shared scientific resources. He will demonstrate the platform through a materials screening pipeline and the execution of modern ML interatomic potentials on HPC systems.

• 9:45 am | Talk 1: Monitoring & Analytics on FABRIC (15 mins)
Speaker: Nik Sultana, Illinois Institute of Technology

Focus: Nik will present Patchwork, an open-source capture and analysis platform designed to run entirely on the FABRIC testbed. Patchwork operates both at the level of individual experiments and across all experiments running simultaneously, without requiring any modifications to the testbed itself. To achieve scale, it leverages FPGA-based NICs and DPDK for high-performance, non-intrusive monitoring. Now in continuous operation for over a year, Patchwork has generated testbed-wide insights into how researchers collectively use FABRIC. The work was published at IMC 2025, where it received the Best Community Contribution award. Key takeaways:
• Techniques for distributed monitoring of federated research infrastructure
• Approaches for analyzing and visualizing large-scale network data
• How Patchwork can be adapted to other testbed or infrastructure environments
• Ongoing work on extensibility and future capabilities

• 10:00 am | Talk 2: Debugging Virtual Networks with Provenance (15 mins)
Speaker: Alexander Wolosewicz, Illinois Institute of Technology

Focus: Alexander will discuss tenant-driven debugging for virtual networks, addressing challenges that arise when services run as tenants on cloud or research testbeds. His work leverages packet histories and provenance to provide effective debugging without requiring knowledge of the underlying physical hardware. He will demonstrate tools built for a national network research testbed, designed to handle exotic traffic patterns, hardware, and protocols commonly used in research environments. Key takeaways: 
• Virtual networks, including research testbeds, can benefit from comprehensive, user-friendly debugging tools. 
• Provenance-driven approaches enable tenant-level visibility and problem resolution without relying on operator-level access. 

• 9:45 am | Talk 1: Deploying CryoSPARC in HPC Environments (15 mins)
Speaker: Heechang Na, Ohio Supercomputer Center

Focus: Heechang will share practical insights on deploying CryoSPARC for cryo-EM data processing in HPC environments. Drawing on experience at the Ohio Supercomputer Center, he will discuss technical challenges such as storage management, GPU utilization, and version updates across more than 15 deployed instances. The session will also explore strategies for scalable, container-based deployments and lessons learned in supporting multiple research groups. Key takeaways: 
• Best practices for deploying and supporting CryoSPARC in HPC
• Lessons learned from real-world deployments at OSC
• Approaches for scalable, containerized HPC workflows

• 10:00 am | Talk 2: Cryo-EM RCD Facilitation Across Campus CI (15 mins)
Speaker: Charles Christoffer, Purdue University

Focus: Charles will discuss the computational infrastructure (CI) requirements for supporting cryo-electron microscopy (cryo-EM) workflows in campus facilities. He will cover challenges from initial data collection through 3D reconstruction, including fast storage, high-speed networking, GPU compute, batch scheduling in shared clusters, and data retention/sharing using platforms like Globus. Attendees will gain practical insights into enabling both on-campus cryo-EM facilities and broader collaborative communities. Key takeaways:
• Infrastructure needs for cryo-EM data collection and processing
• Strategies for GPU scheduling, storage performance, and batch workflows
• Best practices for archiving and sharing data across labs and collaborators

11:20 pm – 11:35 am | Karen Tomko, Ohio Supercomputer Center (15 mins)

• - Title: Sharing Open OnDemand Apps in the Appverse 
• - Focus: Karen will present the Appverse, a community-driven catalog for Open OnDemand (OOD) applications designed to reduce duplication and simplify app discovery, deployment, and reuse across research computing sites. She will discuss the project’s motivation, design goals, and current implementation, and invite feedback and contributions from the Midwest RCD community. Key takeaways:
  • A central platform for finding, deploying, and sharing OOD apps
  • Reducing duplicated development efforts across institutions
  • Engaging the community for feedback and contributions to improve the Appvers

• 11:50 am – 12:05 pm | Brian Kyle, University of Wisconsin-Madison (15 mins)

• Title: Building a Secure HPC for Restricted Data Research
• Focus: Brian will present on what it looked like for the University of Wisconsin to build a high-performance computing environment modeled after NIST 800-223, steps completed to secure computing and complete a formal risk assessment to allow restricted data (i.e. ePHI, HIPAA) workloads within their institution, review ongoing support and future plans within a deny-by-default network firewall, and talk through practical pain points and how they're addressing them. Key takeaways: 
  • Learn what UW-Madison School of Medicine and Public Health does for:
  • Current approaches to HPC platform automation, configuration management
  • Support tools and approaches in a restricted environment
  • GPU/LLM/AI support and infrastructure when working with restricted data

11:50 am – 12:05 pm | Brock Palen, University of Michigan (15 mins)

• Title: Data Ecosystem: Thinking Beyond Storage
• Focus: Brock will explore the features and capabilities of a research data ecosystem that enhance researcher productivity, drawing on ARC’s long-running initiative to make data more actionable and sustainable. The session will highlight strategies for maximizing data value and making the case for institutional investment in research infrastructure. Key takeaways:
  • Understanding the value of research data beyond storage
  • Strategies for sustainable, actionable data management
  • Making the case for institutional investment in data infrastructure

Speakers: Todd Shechter & Winona Snapp-Childs 
Focus: Todd and Winona will provide an update on the NSF-funded CASC workshop initiative aimed at strengthening state-level coordination in RCD. NSF is asking each state to nominate two representatives, one RCD leader and one state or government-level representative with a broad systems view, to help re-envision how CI investments can have impact beyond a single institution. They will discuss the results and takeaways from that meeting.. The session will also address the challenges states face in identifying appropriate representatives and building statewide RCD structures, as well as what this means for the Midwest region. Key takeaways:

• Understanding NSF’s new push for state-level RCD coordination
• What the upcoming CASC/NSF workshop is trying to achieve
• Why NSF wants broader-impact, cross-institutional strategies
• Opportunities for Midwest RCD leaders to engage and shape the conversation

• 2:20 – 2:30 pm  | From Physics to Research Computing: An Early-Career Perspective on Entering RCD (10 mins)
Speaker: Smita Darmora, University of Illinois Chicago

Focus: Smitha will share insights from her transition from an experimental high-energy physicist working on ATLAS at CERN to a computational scientist supporting research computing users. Having experienced both sides, as a domain researcher and now as RCD staff, she highlights the persistent gaps between scientific needs and computational infrastructure. Her talk will cover common challenges STEM researchers face when entering distributed and grid-based computing environments, including software environment issues, workflow translation, and navigating complex systems. She will also discuss which training, documentation, and support approaches were most effective from a researcher’s perspective, and where communication gaps often occur between scientists and RCD teams. Key takeaways: 
• Challenges domain researchers face when adopting distributed computing systems
• Differences between researcher and RCD staff perspectives
• Effective training, onboarding, and support strategies
• Practical ways to lower barriers and improve collaboration for early-career researchers

• 2:30 – 2:40 pm | Developing the Research Computing Workforce Pipeline from K–12 to Early Career through Experiential Learning (10 mins)
Speaker: Amanda Hassenplug, Purdue University

Focus: Amanda will present RCAC’s multi-layered workforce development model, spanning K–12 outreach, the Anvil REU, and the CI-XP student program, all designed to create inclusive, skills-based pathways into research computing and data (RCD) careers. She will highlight how K–12 camps build early awareness of computational thinking and emerging technologies; how the Anvil REU offers immersive advanced cyberinfrastructure research experiences; and how the CI-XP program provides year-round experiential learning in HPC, XR/AR/VR, cyberinfrastructure operations, and outreach. The session will cover program design, recruitment, mentoring, assessment, outcomes, partnerships, lessons learned, and future plans. Key takeaways: 
• Successful workforce development models
• Student engagement from K–12 through early career
• Building experiential learning and mentorship pathways
• Outreach strategies and sustainable funding approaches

• 2:40 – 2:50 pm | Midwest Carpentries Community (10 mins) (CANCELLED)
Speaker: Sarah Stevens, UW-Madison

Focus: Sarah will showcase the initiatives that are advancing training and collaboration across the Midwest Carpentries Community: 
• Midwest Carpentries Community: Ongoing regional efforts to build capacity, support instructors, and strengthen cross-institution collaboration. 
• Building Better Research Software Workshop: A new Midwest Carpentries Community, developed intermediate workshop that helps researchers move beyond the basics into writing sustainable, well-structured, documented, and reproducible software. 
• Collaborative Lesson Development Training: How The Carpentries’ evidence-based framework supports teams in creating high-quality, open-source lessons through structured audience analysis, learning objectives, practice tasks, and assessments. 
• Key Takeaway: A cohesive look at how the Midwest Carpentries Community is closing skills gaps, supporting intermediate learners, and fostering shared lesson development across the region.
• 2:50 pm – 3:00 pm: Q/A for the speakers

Facilitators: Jess Haney & Yan Zhan, NCSA
Focus: Jess and Yan will facilitate an interactive discussion comparing the operational realities of cloud platforms like OpenStack with traditional HPC environments. Drawing on experience across HPC, HTC, Java web services, OpenStack, and Kubernetes, they will highlight how cloud and HPC solve similar problems with very different models, covering rolling upgrades, distributed monitoring, overcommit-based capacity planning, self-service onboarding, and staffing structures. Attendees will walk through real-world scenarios and explore where cloud and HPC operational practices meaningfully diverge, overlap, or complement each other. Outcomes and insights will feed back into the MWRCD Scientific Cloud Affinity Group. Key takeaways:

• How cloud platforms handle day-2 operations differently from HPC
• Tradeoffs between cloud self-service onboarding and HPC allocation-driven access
• Differences in staffing models, workloads, and operational expectations
• Practices that translate effectively between cloud and HPC ecosystems
• Opportunities to connect with the MWRCD Scientific Cloud Affinity Group for continued knowledge-sharing

Facilitator: James Cannon, University of Michigan 
Focus: James will facilitate this BoF to bring together participants from across the Midwest to discuss how institutions are building and sustaining effective support structures for AI in research. Attendees, whether leading established AI consultation programs or just beginning to develop them, will share experiences, strategies, and lessons learned. The conversation will highlight best practices, common challenges, and opportunities for collaboration, with the goal of strengthening the region’s collective capacity to support innovative, AI-enabled research.