Elastic Expert Parallelism in @vllm_project lets you add or remove GPUs from an active Mixture-of-Experts deployment during traffic with minimal interruption to serving and minimal downtime. At #PyTorchCon North America 2026, Itay Alroy (@nvidia) will present “Elastic Expert Parallelism in vLLM,” covering the architecture, key implementation details, open challenges, and future roadmap for Elastic EP. Alroy will also examine what happens when EP size changes and how NIXL EP enables grow/shrink under live traffic. Register for PyTorch Conference North America 2026: hubs.la/Q04v4SL60

Debugging LLM training in production is notoriously challenging because subtle bitwise errors can surface long before they ever trigger a spike in the loss curve. Ziming, a PhD student at the University of Michigan and student researcher at @ByteDanceSeed_ , will speak at PyTorch Conference North America on how to solve this issue using OpGuard. OpGuard compares separate training runs bit by bit to pinpoint the exact initial operation where executions diverge. Join us in San Jose, October 20 to 21, to learn how bitwise alignment delivers faster, more precise LLM training debugging: hubs.la/Q04v4SL60 #PyTorchCon @UMich

Going to @PyTorch Conference in San Jose next month? @CrusoeAI and we are hosting an evening on day one, October 20, 6:30 PM, right after the sessions. We're gathering GPU experts, AI researchers, AI infra engineers, and anyone into AI infra and heterogeneous AI compute. Space is limited: luma.com/6j69xyal

See you there!

Interested in optimizing Mixture of Experts LLM inference on ARM CPUs? In his talk at PyTorch Conference North America, Maajid Khan from Fujitsu Research India will explore efficient MoE LLM inference using vLLM and OpenVINO, sharing practical strategies for running these models effectively on ARM architectures. Register and join us in San Jose, October 20-21: hubs.la/Q04v4SL60 #PyTorchCon

To support developers, researchers, and engineers in deepening their technical skills, the PyTorch Foundation is running a dedicated Introduction Track at PyTorch Conference North America in San Jose, California, featuring the following speakers: @NVIDIA - Aastha Jhunjhunwala, Mark Moyou @RedHat- Lucas Wilkinson, Matthew Bonanni @HuggingFace - Aritra Roy Gosthipaty, Suvaditya Mukherjee @CrusoeAI - JanakiRam Goteti, Suman Debnath ETH Zurich - Andrea Mattia Garavagno @Harvard University - Vijay Janapa Reddi @CloudNativeFdn - Yashasvi Misra Additionally, an official, full-day PyTorch Associate Training session led by instructor Faradawn Yang will take place on Monday, October 19, 2026, offering hands-on labs, real-world projects, and a voucher for the PyTorch Certified Associate exam. Seats are limited, so register by Tuesday, October 13th to secure a spot and elevate your AI development skills. Read our latest blog to learn which sessions will feature at the Introduction track and how to sign up to the PyTorch Associate Training session: bit.ly/4dE9899

Join us October 17-18 in San Francisco for the ExecuTorch Hackathon. Developers will build and deploy PyTorch models with ExecuTorch across Compute, Mobile + XR, and IoT tracks. Teams will get hands-on with Snapdragon-powered PCs, Samsung Galaxy devices, and Arduino hardware while exploring what's possible with open-source, on-device AI. Form a team, submit a proposal, and help shape the future of AI at the edge. Submit your proposal here: luma.com/executorch-hackatho…

At #PyTorchCon North America 2026, @lizli202503 and @indianspeedster of @AMD will discuss their work enabling scalable MXFP8 pretraining on MI355X GPUs. Liz and Shekhar will cover optimization of key MXFP8 operators in TorchAO and how those kernels were integrated with TorchTitan to build an end-to-end upstream PyTorch training stack for large language models. Their session, “Scaling MXFP8 Pretraining on 1K+ AMD Instinct MI355X: TorchAO Kernels and TorchTitan Training,” will also cover Triton and FlyDSL implementations and lessons from numerical validation, convergence, and performance tuning. Join us in San Jose on October 20-21. Register today: hubs.la/Q04v4SL60

🤲 Open source moves forward when people build together. #PyTorchCon North America brings together the researchers, engineers, maintainers, and contributors doing exactly that - sharing what they've learned, working through challenges, and building on each other's work. Join the open source AI community in San Jose from Oct 20-21: bit.ly/4sh3DSw

🚨 Last Chance To Take The Stage At PyTorch Day Japan! The Call for Proposals closes this Sunday, September 27 at 11:59 PM JST. Bring your ideas and experiences to a community of PyTorch developers, machine learning engineers, AI researchers, and open source AI leaders from across Japan and beyond. ⏰ Submit Your Proposal >> bit.ly/4yuIMyz

At #PyTorchCon North America 2026, Alessandro Sangiorgi, Senior Software Engineer (@RedHat), will show how warm starts let Helion autotuning reuse cached configurations so developers can iterate faster instead of repeating cold searches. Sangiorgi will present “Accelerating Helion Autotuning with Warm-Start and Shared Caches,” a poster on Helion, a PyTorch DSL for GPU kernels. Join us in San Jose on October 20-21: events.linuxfoundation.org/p…

The PyTorch Foundation OSPO & Academic Outreach Working Group is inviting academic open source projects to submit for a Day 0 workshop at #PyTorchCon North America 2026. Projects using PyTorch, @DeepSpeedAI, Helion, @raydistributed, Safetensors, or @vllm_project can submit for consideration. Eight selected projects will present five-minute lightning talks, followed by a hands-on workshop and mentor office hours. Depending on the project, workshop conversations may cover open source readiness, governance, reproducibility, community, security and maintenance, and ecosystem positioning. The call is open to students, researchers, faculty, research engineers, university labs, academic OSPOs, research institutions, and maintainers of research-originated open source projects. The submission deadline is September 30, 2026, at midnight PT. Register for PyTorch Conference North America: hubs.la/Q04tBgv_0 🔗 Read more and submit: pytorch.org/blog/from-resear…

This just in! Take our live, in-person PyTorch Associate Training the day before PyTorch Conference North America in San Jose, CA on Monday, October 19th. Through a blend of lectures, demonstrations, guided labs, and assessments, you'll will not only become more familiar with PyTorch’s ecosystem but will also build and optimize real-world models from scratch. Upon completion of the course, you will also receive a voucher ($250 value) to take the new PyTorch Certified Associate (PTCA) certification exam. Learn more and sign up for this training at: hubs.la/Q04y8msv0 #PyTorchCon

Following a great community gathering at Graphcore's London office, momentum is building for PyTorch Conference North America 2026. Join PyTorch Foundation board member Sylvain Viguier and fellow open source developers, researchers, and technical leaders in San Jose, California on October 20 and 21. We look forward to hearing about your latest projects, sharing technical insights, and connecting in person. Register now to save your spot: hubs.la/Q04v4SL60 #PyTorchCon @graphcoreai

Build AI on the edge with ExecuTorch at a two-day @Meta + @Arm hackathon at San Francisco State University. Students, developers, and AI enthusiasts can get hands-on experience building AI applications with ExecuTorch on Arm-powered devices, with technical talks, demos, workshops, and mentorship. Participants will also have the chance to collaborate with fellow builders, showcase their projects to judges from Arm and Meta, and compete for prizes and recognition. 🔗 Learn more and register: arm-executorch.devpost.com/?…

Open software. Open hardware. Open standards. Open data. Open benchmarks. If your Bay Area event advances open AI, it could be part of #OpenSourceAIWeek, October 16-25, featuring flagship events #PyTorchCon and #AGNTCon + #MCPCon. Submit to be a part of it all: bit.ly/4ipYZ39

How do you keep @vllm_project moving at the speed of light without excluding users who run diverse models on diverse hardware? In a new PyTorch Foundation blog, contributors from @IBM, @Meta, and @huggingface introduce hardware-agnostic layers designed to balance frontier performance with portability, helping ensure vLLM continues to meet the needs of the broader open source ecosystem. Read the blog to learn more: bit.ly/4xEH7Wk @hmellor_ @th_ortner

Where research meets real-world application. 🌎 At #PyTorchCon North America, hear from the researchers and practitioners advancing #PyTorch and putting those advances to work. See how ideas move from research into practical implementation and how what gets learned in practice helps shape what comes next. Join us in San Jose, Oct 20-21: bit.ly/4sh3DSw

PyTorch releases are no longer just about shipping PyTorch binaries. Release engineering now coordinates PyTorch, Triton, and @vllm_project as part of a broader ecosystem. At #PyTorchCon North America 2026, Andrey Talman (@Meta) will discuss how the PyTorch release process is being modernized, including the use of AI agentic workflows to analyze failures and the work required to release projects together with reliability and performance in mind. Hear more from Andrey about what he’ll cover at the conference. 🔗 Register: events.linuxfoundation.org/p…

Article by @cmazzaanthony & @Drewch + @cmazzaanthony will keynote at #PyTorchCon NA on the flywheel that turns Shopify's product expertise into robust evaluations, then mines, critiques, and repairs low-scoring conversations to train better models.