PyTorch, Powering the Enterprise
PyTorch, Powering the Enterprise
brought to you by Red Hat, NVIDIA, and IBM
PyTorch, and its ecosystem, such as vLLM, Ray, and Helion, have grown to become the foundation of the AI world. Along with this growth PyTorch has moved from the research labs to power the largest Enterprise companies. Come learn and interact with the companies helping to bring PyTorch to its full enterprise potential, hosted by Red Hat, NVIDIA, and IBM.
Agenda
9:30–9:45 AM — PyTorch and the Open Ecosystem Behind Enterprise AI
Speaker: Stephen Watt (PyTorch Board Member, Red Hat)
How does an open source research framework become a foundation for enterprise AI? This opening keynote explores the role of PyTorch, its community and the broader ecosystem in connecting model development with real-world applications. Attendees will learn how shared infrastructure, open collaboration and customer participation shape the choices available to businesses investing in AI.
9:45–10:15 AM — From Models to Business Value: Navigating the PyTorch Ecosystem
Speakers: Joseph Groenenboom (PyTorch Ecosystem and TAC, Red Hat), Chris Hoge (Chair PyTorch Ecosystem, NVIDIA)
A useful model is only one part of a successful AI application. This session begins with the core PyTorch project and maps the journey from model development to deployment and ongoing operations, showing where PyTorch and complementary ecosystem projects fit. Business and technology leaders will leave with a framework for evaluating interoperability, skills, support and operating costs—and deciding which capabilities to build, adopt or source from partners.
10:15–10:45 AM — The Economics of Serving AI: What Business Leaders Should Know About vLLM
Speaker: Tyler Michael Smith (vLLM Core Maintainer, Red Hat)
As AI usage grows, serving models becomes a business decision about cost, responsiveness and capacity. This session explains where vLLM fits in the ecosystem around PyTorch and how serving software and hardware choices affect those tradeoffs. Attendees will learn which questions to ask about cost per request, service quality and scaling, and how to assess ecosystem advances against the needs of their own applications.
10:45–11:00 AM — Morning Break
11:00–11:30 AM — Scaling AI Services: A Customer's Journey to Production Inference (llm-d)
Speakers: Maroon Ayoub (llm-d Maintainer, Red Hat)
What changes when an AI pilot becomes a service that people depend on? This customer case study will examine the business need, platform decisions and operational lessons behind scaling inference, with llm-d as the proposed ecosystem lens. Attendees will learn how to assess reliability, capacity and operational complexity, and which results to measure when deciding whether a distributed approach is worth the investment.
11:30 AM–12:00 PM — Making AI Pay: A Customer Case Study in Inference Economics
Speakers: Kyle Kranen (Dynamo Maintainer, NVIDIA)
How do the economics of inference shape what a business can offer its customers? This case study will follow an AI application from its business objective through the choices made to serve it at scale, with NVIDIA Dynamo as the proposed ecosystem lens. The discussion will connect cost, latency and resource use to customer experience and commercial goals, giving attendees a practical way to evaluate their own AI investments.
12:00–12:30 PM — Agents and Inference Panel
Moderator: Ankit Patel (PyTorch Board Member, NVIDIA)
Bring the morning's ecosystem and customer discussions together around three decisions: which business problem to pursue, how to measure success and what capabilities are needed to deliver it. Attendees will leave with questions to take back to their teams and into the afternoon's conversations about production platforms and AI agents.
12:30–1:30 PM — Lunch and Peer Networking
1:30–2:00 PM — Path to Production AI
Speakers: Christopher Nuland (llm-d contributor, Red Hat), Babak Mozaffari (NVIDIA)
Moving one model into production is different from supporting AI across a business. This customer-led discussion will examine how teams establish shared infrastructure, operational ownership and governance, using the collaboration between Red Hat and NVIDIA as a concrete platform example. Attendees will learn how to assess a common foundation for PyTorch-based workloads and other AI applications, what to standardize and how to measure delivery speed and operating effort.
2:00–2:30 PM — AI Agents at Work: Balancing Business Value, Control and Trust
Speakers: Kris Murphy (NVIDIA), Sally O'Malley (OpenClaw Maintainer, Red Hat), Mrunal Patel (OpenShell Maintainer, Red Hat)
Giving an AI agent access to business tools creates new opportunities—and new operating decisions. This talk will examine a proposed workflow, the permissions and controls it requires, and the evidence needed before expanding its use. OpenShell, OpenClaw and related ecosystem work will provide context for the discussion. Attendees will leave with a framework for evaluating usefulness, oversight and operational readiness, from the underlying models to the actions an agent can take.
2:30–3:30 PM — Closing Reflections and Networking Coffee
Speaker: Ankit Patel (PyTorch Board Member, NVIDIA)
Reflect on how PyTorch and its ecosystem connect AI innovation with business applications and identify a practical next step for your organization. Continue the discussion over coffee with customers and contributors, compare lessons from the day and explore opportunities to collaborate.
Please register with your business email address. Space is limited to 50, so sign up now to reserve your seat!
Privacy Notice: The information we collect in relation to the event will be processed in accordance with Red Hat's Privacy Statement. As detailed in the Privacy Statement, Red Hat may use your information to provide you with information about the event and may share your information with service providers that provide services related to the event. Providing your information to Red Hat may involve transferring it to another country.