Name: KubeCon + CloudNativeCon North America
Start: 2026-11-09
End: 2026-11-12

Michael Amundson

Distinguished Engineer

CVS Health

Michael Amundson is a distinguished engineer in the Developer Experience organization at CVS Health where he leads technical direction for the CAP Application Platform. His near decade of experience in platform engineering started at Target Corporation, where he helped build Target's in-house application platform. Prior to his work on platforms, Michael worked as an application engineer in roles across a number of Fortune 500 companies, daydreaming of platforms that didn't yet exist.

Conference Sessions

Beyond End User: CAP and the Evolution of Open Source at CVS Health Nov 11, 2026 • 10:23 AM-10:28 AM Salt Palace | Level 1 | Hall DE
Two years ago, the DevEx organization at CVS Health set an audacious plan in motion: build a platform capable of running the company’s thousands of existing applications while remaining approachable enough that someone joining the company could ship a new application on their very first day. To make this work, we knew we would need to leverage many of the incredible open source projects across the CNCF landscape. But we also knew that achieving our goal would require software we’d have to build: a meta control plane that could weave these projects together into a single streamlined platform. We call the platform CAP and we’re excited to share its design, show a brief demo, and talk about our transition from end user to open source maintainer and contributor as CAP evolves from an internal platform to a brand new open source project.

Chris Aniszczyk

CTO, Cloud and Infrastructure

The Linux Foundation

building a better world through open collaboration

Conference Sessions

Welcome + Opening Remarks Nov 10, 2026 • 9:00 AM-9:30 AM Salt Palace | Level 1 | Hall DE

Sunil Arora

Software Engineer

Google

Sunil Arora is a software engineer on the GKE team at Google Cloud and a long-time Kubernetes contributor. He was an original author and lead maintainer of Kubebuilder, has made foundational contributions to kubectl, kustomize, and kpt, and co-authored kubectl-ai. He now leads OpenRL, Kubernetes-native post-training infrastructure for LLMs. For over a decade, Sunil has built tools that help developers and platform teams safely deploy and manage applications at scale.

Conference Sessions

Pods in the Wild: Reinforcement Learning on Kubernetes with OpenRL Nov 10, 2026 • 9:32 AM-9:40 AM Salt Palace | Level 1 | Hall DE
What do wild dolphins, open-weight Gemma models, and Kubernetes have in common? Pods, of course! By training DolphinGemma on decades of marine acoustic data, researchers are actively decoding dolphin communication. This reflects a broader industry shift: transformative AI pairs domain-specific open weights with specialized data. Far from replacing frontier models, winning architectures unite massive foundation LLMs for broad reasoning with lightweight, task-tuned models for targeted execution. While frontier models set the bar for general intelligence, the next major operational leap belongs to everyday builders: empowering practitioners to fine-tune open-weight models that outperform massive frontier baselines on domain-specific tasks at a fraction of the cost, footprint, and latency. Historically, reinforcement learning (RL) was trapped inside elite research labs and brittle, complex runtimes. A new CNCF sandbox project, OpenRL fundamentally changes this paradigm by democratizing RL directly on Kubernetes. Now, a builder can articulate a task, point to a dataset, and use CNCF technologies to guide an open-weight model toward task mastery without requiring deep machine learning expertise.

Keith Babo

Chief Product Officer

Solo.io

Keith Babo is the Chief Product Officer at Solo.io, where he leads product strategy and execution, enabling enterprises to build and scale the next generation of cloud-native and agentic infrastructure for AI-driven applications. Prior to Solo.io, Keith held leadership roles in product management and engineering at Red Hat, Sun Microsystems, and Intel Corporation.

Conference Sessions

The Harness Is the New Runtime Nov 10, 2026 • 10:12 AM-10:17 AM Salt Palace | Level 1 | Hall DE
Agent harnesses are the dominant pattern for agentic AI on the desktop and will become the dominant runtime for agents on Kubernetes. The hard part is the move: bringing the harness onto your cluster with end user experience intact while adopting Kubernetes primitives for sandboxed execution, checkpoint/restore, scale, and security. Containers made this exact trip from developer machines to shared infrastructure in Kubernetes. In this keynote we'll present harnessed agents as your cluster's next critical workload. We'll show how harnesses combine models with a modular runtime of tools, skills, and plugins, and how leading CNCF ecosystem projects deliver the identity, policy, and observability to run them in production.

Kathryn Baldauf

Software Engineer

Apple

Kathryn is a software engineer at Apple, where she works on Apple's open source Container and Containerization projects. She has been working in the open source container ecosystem for the past 7 years.

LinkedIn

Conference Sessions

Code, Containers, and Clusters: The Next Chapter of Apple's Container Framework Nov 11, 2026 • 9:38 AM-9:48 AM Salt Palace | Level 1 | Hall DE
Run Linux containers on macOS. Build with Apple silicon. Deploy at scale. Last year at KubeCon, we introduced Apple’s open source container framework. We kept building. Container Machine: The newest addition to the framework brings new functionality to Apple silicon. Container machines provide a highly integrated Linux environment that works seamlessly on your Mac. It’s fast, lightweight, and persistent. Kubernetes Plugin: We’re previewing a plugin for running Kubernetes locally on Apple silicon. Provision high-fidelity local clusters using upstream-native tools, such as kubeadm, kind images, and Cilium. At Apple, containers run at massive scale, while local testing maintains reliability. Watch us code a Kubernetes controller, then deploy it to a local cluster spun up with the plugin. All on Apple silicon. Leave with the roadmap, hands-on experience, and an invitation to join the community.
BoF | Let’s Discuss: The Future of Apple Containers with Cloud Native Tooling Nov 11, 2026 • 11:50 AM-12:20 PM Salt Palace | Level 2 | 254
Apple's open source Containerization framework lets developers run Linux containers on macOS in lightweight virtual machines. With new support for Container Machines and an experimental plug-in for local Kubernetes clusters, the project brings native, secure, and fast cloud tooling to developers with a Mac. Join this session to meet some of the development team, share how you use containers today, offer feedback on the project’s roadmap, and discuss best practices for using it as part of the cloud native ecosystem.

Abby Bangser

Principal Engineer

Syntasso

Abby is a Principal Engineer at Syntasso delivering Kratix, the open-source cloud-native framework for providing the execution later for your internal platforms. Her keen interest in supporting internal development comes from over a decade of experience in consulting and product delivery roles across platform, site reliability, and quality engineering. Abby is a CNCF Ambassador, Team Topologies Advocate, emeritus CNCF Platforms Working Group lead, and co-host of the #CoffeeOps London meetup.

Conference Sessions

The Flywheel Effect: Why Community Is the Ultimate End-User Advantage Nov 11, 2026 • 9:00 AM-9:10 AM Salt Palace | Level 1 | Hall DE
Cloud native promises a future without limits, but every end-user organization eventually discovers the same truth: no team becomes limitless alone. Every day, end users push CNCF projects into new industries, at new scales, and with entirely new challenges. They uncover new patterns, expose new edge cases, and find better ways to build. By sharing those experiences back with the community through talks, bug reports, reference architectures, and open collaboration, they create a flywheel where every contribution accelerates the next innovation. The result is a community where shared experience helps everyone solve problems faster, see around corners, and innovate with greater confidence. Together, Phill Morton and Abby Bangser will explore how end-user experience and community leadership fuel that flywheel. Through the upcoming Platform Engineering Community White Paper and a live demonstration of an enterprise AI platform built on those same principles, they’ll show how community ideas translate into real-world impact, enabling thousands of non-engineers to build at the speed of agents. You’ll leave with practical ideas for applying those patterns in your own organization and a new perspective on what “Cloud Native Without Limits” really means: the greatest advantage in cloud native isn’t just the technology we build; it’s the community we build it with.

Warren Barkley

Vice President of Product Management

NVIDIA

Conference Sessions

Building the Next Generation of AI with Open Source Nov 10, 2026 • 9:49 AM-9:54 AM Salt Palace | Level 1 | Hall DE
Join us to hear how NVIDIA is deepening its commitment to open source and working alongside the Kubernetes community to build the future of AI Factories—together. We’ll share how an open, community-driven approach is helping make accelerated computing easier to build, operate, and scale.

Dmitry Berkovich

Software Engineer

Google

.

Conference Sessions

Agent Substrate: Open Layers for Agentic Systems Nov 10, 2026 • 10:03 AM-10:10 AM Salt Palace | Level 1 | Hall DE
Running agentic workloads on Kubernetes presents several challenges. Agent Substrate is a new project, built on and with Kubernetes, which enables you to securely run huge numbers of agents on Kubernetes with ultra low latency.
Kubernetes Solutions for Agent-Shaped Problems Nov 10, 2026 • 4:40 PM-5:10 PM Salt Palace | Level 1 | Hall DE
Agents and agent-like workloads are becoming a predominant class of workloads on Kubernetes, but they are different from "traditional" apps. In the last 18 months, those differences have been starkly emphasized. For over a decade K8s has been the industry standard platform for workloads of all sorts, with a rich ecosystem that has successfully adapted to every new class of workload. It’s time to adapt again. The Agent Substrate (http://ate.dev) project was born to tackle the problem of agent-like workloads head-on. Launched in May, Agent Substrate’s primary goal is to make it easy and efficient to run agent-like workloads. It leverages Kubernetes for what it is best at, then builds up from there to address the particular needs of agents. A lightweight control-plane coupled with active session management and network-triggered wake-ups enables users to run orders of magnitude more agent-like workloads than before – more efficiently and with lower latency and higher throughput.

Jim Beyers

Vice President

CVS Health

Hey there! I'm Jim on the CVS DevEx team, blending my love for tech with making life easier for developers. After stints at Target and Capital One, I've learned to connect tech and business. Originally from Wisconsin, I landed in Minneapolis and collected degrees like Pokémon cards (BA in CS, MS in Software Engineering, MBA from St. Thomas). Off hours, I'm with my wife Nora, kids Amelia and Ben, and our husky Cali Bean. Big on music, concerts, films, and Age of Empires 2.

Conference Sessions

Beyond End User: CAP and the Evolution of Open Source at CVS Health Nov 11, 2026 • 10:23 AM-10:28 AM Salt Palace | Level 1 | Hall DE
Two years ago, the DevEx organization at CVS Health set an audacious plan in motion: build a platform capable of running the company’s thousands of existing applications while remaining approachable enough that someone joining the company could ship a new application on their very first day. To make this work, we knew we would need to leverage many of the incredible open source projects across the CNCF landscape. But we also knew that achieving our goal would require software we’d have to build: a meta control plane that could weave these projects together into a single streamlined platform. We call the platform CAP and we’re excited to share its design, show a brief demo, and talk about our transition from end user to open source maintainer and contributor as CAP evolves from an internal platform to a brand new open source project.

Lorenzo Biava

Staff Infrastructure Engineer

LiveKit

Lorenzo Biava a Staff Infrastructure Engineer at LiveKit, where he operates a global WebRTC network pushing over 10 PB of monthly real-time traffic across dozens of regions. He previously served as Principal Engineer at Microsoft and Nuance, building secure, globally-distributed Kubernetes platforms for Voice AI in strictly regulated environments.

Conference Sessions

Production AI on Kubernetes: Built to Scale with Oracle Cloud Nov 11, 2026 • 10:04 AM-10:09 AM Salt Palace | Level 1 | Hall DE
Kubernetes is the foundation of the next generation of AI. Moving from experimentation to a production requires infrastructure designed for scale, performance and simplicity. In this keynote, Oracle Cloud Infrastructure (OCI) will explore how OCI Kubernetes Engine (OKE) enables organization to build and run demanding cloud-native and AI/ML workloads on an open, enterprise-ready platform. Join Jeff Hoffman, VP of Container Services at Oracle, to hear how customers are using OKE to bring AI initiatives into production. Learn how OCI’s Kubernetes platform, high-performance infrastructure, and open-source ecosystem help teams innovate faster while maintaining the control and reliability their critical workloads require.

Dilpreet Bindra

Sr. Director Engineering, VCF Division

Broadcom

Dilpreet Bindra is the Senior Director of Engineering for the VMware Cloud Foundation (VCF) division at Broadcom, where he leads the delivery of runtime platforms, automation, and modern cloud-native architectures. He is a passionate advocate for bridging the gap between developer agility and enterprise IT governance, with a current focus on integrating Kubernetes and Private AI infrastructure into enterprise environments. A seasoned engineering leader, Dilpreet’s expertise spans both hardware and software systems. Prior to his long tenure shaping core VMware technologies—including ESXi and vCenter—he held software and hardware engineering roles at Sun Microsystems and HP. Within the open-source community, he actively champions Broadcom’s Platinum membership in the CNCF and drives continuous upstream collaboration. Dilpreet holds a Bachelor of Science in Electrical Engineering from Stony Brook University. Outside of his engineering work, he is a basketball player, hiker, pickleball player, and dedicated father.

Conference Sessions

One Kubernetes, Different Environments: Closing the Portability Gaps Nov 10, 2026 • 9:56 AM-10:01 AM Salt Palace | Level 1 | Hall DE
What does portability really mean when Kubernetes spans containers, virtual machines, and increasingly diverse infrastructure? This keynote will discuss how VM API and the Workload Conformance Working Group are tackling different sides of that challenge and building more consistent infrastructure APIs and workload expectations across the Kubernetes ecosystem.

Vara Bonthu

Principal Open Source Specialist SA

AWS

Vara Bonthu is a Principal Technologist at AWS and CNCF maintainer for the Kubeflow Spark Operator and MCP Server for Spark History Server. He leads the Data-on-EKS and AI-on-EKS open-source initiatives and has operated Spark on Kubernetes at the scale of thousands of nodes and hundreds of thousands of pods. His work focuses on performance, resiliency, and AI-driven observability for Spark, Ray, and Kubeflow.

Conference Sessions

Why We Bet the Data Platform on Kubernetes – And How AI Is Changing How We Operate It Nov 10, 2026 • 11:30 AM-12:00 PM Salt Palace | Level 1 | 151
Kubernetes has quietly become the default substrate for petabyte-scale data processing. Salesforce, Pinterest, Airbnb, and Apple all run native Spark, Flink, Trino, StarRocks, and Iceberg on Kubernetes at a scale the platform was never designed for. This panel of end-user operators talks about two things data leaders keep asking about. First, what does it actually take to run large-scale data on Kubernetes? Shuffle bottlenecks and Apache Celeborn, stateful upgrades, multi-tenant scheduling with YuniKorn and Kueue, Iceberg storage layouts, mixed CPU and GPU. Each panelist shares what broke at scale and how they fixed it. Second, how are AI agents changing day-2 operations? Troubleshooting agents for Spark and Flink, MCP servers for Spark History Server, agentic upgrade planning, autonomous tuning, AI-driven governance and lineage. Where are teams seeing real productivity gains, and where do agents still get it wrong? Production stories, real disagreements, no vendor pitches.
End-User Panel – KubeFlow Enterprise Use Cases-Solutions Nov 11, 2026 • 2:08 PM-2:28 PM Salt Palace | Level 1 | Solutions Showcase | Halls 5 | Cosmic Canyon Zone | Cloud Native Theater
Panelists will discuss how their organizations leverage Kubeflow to orchestrate AI workloads on Kubernetes—from distributed training to low-latency serving—while overcoming enterprise challenges such as scalability, pipeline automation and hybrid-cloud integration. Attendees will gain actionable insights into architecture design, lessons learned from scaling AI workloads on Kubernetes, and proven patterns for accelerating model deployment from experimentation to production.
Kubeflow Graduation: Powering the Next Era of Cloud Native AI With an Open Community Nov 12, 2026 • 9:12 AM-9:17 AM Salt Palace | Level 1 | Hall DE
Since 2017, Kubeflow has evolved from a TensorFlow controller into a Cloud Native AI platform comprising eleven full-fledged projects, used by thousands of organizations across the AI lifecycle. Today, we are excited to celebrate Kubeflow’s Graduation from the CNCF – and the open community that made this journey possible. This keynote explores real-world stories from our users, showing how an open, vendor-neutral community is pushing the boundaries of Data & AI infrastructure across the whole AI lifecycle. It will highlight large-scale distributed training with Kubeflow Trainer and MPI, and look ahead to the agentic era with innovations such as the Semantic Operator. By bringing governed semantic APIs to Kubernetes, the Semantic Operator enables AI agents, applications, and BI tools to work with trusted business definitions without relying on LLM-generated SQL. Together, these efforts illustrate how the Kubeflow community is building the foundations for the next decade of Cloud Native AI.

Jonathan Bryce

Executive Director, Cloud and Infrastructure

The Linux Foundation

Jonathan Bryce is the Executive Director of Cloud & Infrastructure at the Linux Foundation, where he leads both the Cloud Native Computing Foundation (CNCF) and the OpenInfra Foundation—two of the largest and most influential open source communities in the world. With over 15 years of open source leadership, Jonathan co-founded the OpenStack project, guiding its evolution into a global ecosystem that supports a multi-billion-dollar market and serves as the de facto standard for open cloud infrastructure. Before his open source career, Jonathan spent over a decade in data center infrastructure and co-founded one of the industry’s earliest cloud computing companies, which grew into a billion-dollar business following its acquisition by Rackspace. Passionate about open collaboration, infrastructure innovation, and empowering users, contributors, and ecosystems worldwide, Jonathan frequently speaks at international conferences on open source strategy, infrastructure modernization, cloud-native technologies, and open source AI initiatives.

Conference Sessions

Welcome + Opening Remarks Nov 10, 2026 • 9:00 AM-9:30 AM Salt Palace | Level 1 | Hall DE

Jorge Castro

Cloud Native Computing Foundation

Conference Sessions

Celebrating the People Who Make Cloud Native Possible Nov 12, 2026 • 9:19 AM-9:30 AM Salt Palace | Level 1 | Hall DE
Behind every project, release and community milestone are people doing the work that keeps cloud native moving forward. The CNCF Community Awards recognize individuals whose technical contributions, sustained leadership and often unseen work have made an exceptional impact on the cloud native community. Join us as we celebrate community members across four awards, recognizing the many ways people contribute to building and sustaining cloud native technologies.

Kai-Hsun Chen

Infrastructure Lead

Hark

Kai-Hsun leads infrastructure at Hark, spanning compute, data, and the full model lifecycle. Previously, he worked on infrastructure at xAI and Anyscale. He is also an open-source enthusiast who maintains several projects, including Ray, KubeRay, and Apache Submarine, and has been contributing actively to Kueue over the past few months.

Conference Sessions

KubeRay 101 Nov 11, 2026 • 11:00 AM-11:30 AM Salt Palace | Level 2 | 254
KubeRay is the Kubernetes-native operator for running Ray workloads, providing lifecycle management for Ray clusters, jobs, and production services on Kubernetes. This Birds of a Feather session is an introduction to KubeRay led by members of the KubeRay Technical Steering Committee. We’ll cover the fundamentals of how KubeRay works, including its core RayCluster, RayJob, and RayService custom resources; common patterns for deploying and autoscaling distributed AI and machine learning workloads; and how KubeRay fits into the broader Kubernetes ecosystem. The session will also be an opportunity to meet KubeRay maintainers, ask questions, discuss real-world deployment challenges, and learn how to get involved in the community. With KubeRay currently applying for donation to the CNCF, we’ll also discuss the project’s community and governance and what CNCF stewardship means for its next phase.
Scaling Multi-Modal Models and Computer-Use Agents on Kubernetes Nov 12, 2026 • 9:32 AM-9:37 AM Salt Palace | Level 1 | Hall DE
Hark builds multi-modal models and computer-use agents — models that operate software the way a person does. Scaling them means running every kind of workload on the same infrastructure: multi-modal data processing, training, reinforcement learning, and inference serving. All of it runs on Ray, deployed and autoscaled by KubeRay. But Ray is not only our runtime — it is also our resource management layer, with MultiKueue governing GPU capacity across every cluster we operate. This keynote covers how that fits together.

Kevin Cochrane

CMO

Vultr

Kevin Cochrane is a 25+ year pioneer of the digital experience space. Kevin co-founded his first start-up, Interwoven, in 1996, pioneered open source content management at Alfresco in 2006, and built a global leader in digital experience management as CMO of Day Software and later Adobe. Kevin has also held senior executive positions at OpenText, Bloomreach, and SAP. Now at Vultr, Kevin is working to build Vultr's global brand presence as a leader in the independent cloud platform market.

Conference Sessions

Welcome + Opening Remarks Nov 10, 2026 • 9:00 AM-9:30 AM Salt Palace | Level 1 | Hall DE

Jaana Dogan

Software Engineer

Google

Jaana is one of the tech leads leading Google's ML/AI systems, specifically focusing on agentic serving and distributed runtimes, including new storage and compute infrastructure.

X
Blog

Conference Sessions

Agent Substrate: Open Layers for Agentic Systems Nov 10, 2026 • 10:03 AM-10:10 AM Salt Palace | Level 1 | Hall DE
Running agentic workloads on Kubernetes presents several challenges. Agent Substrate is a new project, built on and with Kubernetes, which enables you to securely run huge numbers of agents on Kubernetes with ultra low latency.

Jessica Forrester

Senior Distinguished Engineer

Red Hat

Jessica Forrester is a Senior Distinguished Engineer and the Chief Architect for AI Catalyst Engineering at Red Hat, working at the intersection of AI and cloud native as new technologies move into the Red Hat AI product portfolio. She was an early contributor to Red Hat OpenShift, Kubernetes, and their related communities. She has led many cross-functional engineering teams at Red Hat across AI, hybrid cloud services, and Kubernetes / OpenShift Container Platform.

Conference Sessions

Sandbox Your Agents Nov 10, 2026 • 10:19 AM-10:24 AM Salt Palace | Level 1 | Hall DE
AI Agents offer unequalled automation tooling for rapid software development, but with unfettered power comes great headaches, namely how to scope and secure them in cloud native stacks. A problematic agent with too much access, either through choice or a lack of knowledge, can take down entire production stacks and wipe databases. In this talk, we explore how using core community building blocks like Confidential Containers and OpenShell, alongside emerging technology across the cloud native ecosystem, allows anyone to isolate their agentic workloads and software development processes at scale with secure, containerized sandboxes.

Deepak Goel

CTO Cloud Native

Nutanix Inc

Deepak is a technology leader specializing in cloud-native platforms, Kubernetes, distributed systems, and platform engineering. As CTO for Cloud Native at Nutanix, he leads cloud-native strategy, platform architecture, and modern infrastructure initiatives. He has helped drive Kubernetes adoption, scalable data services, and the development of KUDO, later donated to the CNCF, enabling enterprises to modernize applications and operate reliably at scale across hybrid and multi-cloud environments.

Conference Sessions

Integrating Envoy AI Gateway and Kyverno Authz for Realtime Governance of LLM and MCP Traffic Nov 10, 2026 • 4:30 PM-4:55 PM Salt Palace | Level 1 | Solutions Showcase | Halls 5 | Cosmic Canyon Zone | Cloud Native Theater
Agentic AI has evolved models from passive text generators into active operators that invoke APIs, query databases, and execute tool calls. Standard Kubernetes RBAC cannot govern non-deterministic actions driven by unvalidated prompts. This session presents a cloud-native architecture that integrates the Envoy AI Gateway with Kyverno’s declarative policy model to govern autonomous AI intent before execution. The session demonstrates how Envoy acts as a specialized data plane for LLM traffic, using Envoy’s External Processing API to "de-frame" Model Context Protocol (MCP) payloads. The Kyverno service then intercepts these extracted intents, applying declarative, YAML-based policies to authorize or block model actions in real time. Attendees will explore techniques to inspect MCP payloads at the gateway layer, author policy rules for AI agents, and enforce zero-trust boundaries without altering underlying agent code.
GitOps Meltdown: How a Commit Triggered a Cascading Delete of 39 Clusters in Production Nov 10, 2026 • 5:30 PM-6:00 PM Salt Palace | Level 2 | 255 BC
GitOps assumes cluster state converges to Git.But, what happens when the desired state is wrong? This session examines a real incident where a routine Flux reconciliation turned a small git change into a cascading prune across a Kubernetes management platform managing 39 workload clusters. A missing Kustomization triggered inventory deletion, causing Flux controller to delete itself during reconciliation. Resulting in orphaned resources, stuck finalizers, and loss of critical Git repository storage. Using timelines and incident evidence, we trace the failure chain and explore Flux inventory tracking, pruning, finalizers, StatefulSet PVC ownership, CSI reclaim policies, and deletion ordering. We also cover recovery, restoring resources, rebuilding the GitOps control plane, and safely resuming reconciliation. Attendees will learn practical ways to reduce GitOps blast radius, protect stateful infrastructure, and build platforms that remain recoverable when automation fails.
Standardized Yet Adaptable: How Nutanix is Evolving Cluster API with CAREN Nov 12, 2026 • 9:39 AM-9:44 AM Salt Palace | Level 1 | Hall DE
Cloud native innovation has always been driven by the community. From Mesosphere's role as a CNCF founding member to D2iQ's continued Kubernetes leadership and now Nutanix's active contributions across multiple CNCF projects, the mission has remained the same: build together in the open.In this session, we'll revisit that journey and explore how those experiences led to CAREN (Cluster API Runtime Extensions – Nutanix), an extensible framework that enables dynamic cluster configuration and lifecycle management while preserving the simplicity and consistency of Cluster API's declarative model. As we look to the future of Kubernetes fleet management, standardization via Cluster API (CAPI) and the ClusterClass feature has brought immense consistency to cluster provisioning. However, real-world enterprise deployments often hit a wall: they require dynamic customization that strict standardization struggles to accommodate.CAREN bridges the critical gap between strict declarative blueprints and real-world adaptability. By leveraging CAPI's runtime mutation hooks, CAREN dynamically configures and manages Kubernetes clusters without ever needing to fork or touch the base ClusterClass. It empowers platform engineering teams to deliver standardized, yet infinitely adaptable Kubernetes clusters.

Tim Hockin

Distinguished Engineer

Google

Tim has spent most of his career at Google, where he works on Kubernetes and Google Kubernetes Engine (GKE). He is one of the technical leads of the Kubernetes project, and has been part of it since before it was publicly announced. He mostly pays attention to topics like APIs, networking, storage, nodes, multi-cluster, resource isolation, and cluster sharing.

Sessionize

Conference Sessions

Agent Substrate: Open Layers for Agentic Systems Nov 10, 2026 • 10:03 AM-10:10 AM Salt Palace | Level 1 | Hall DE
Running agentic workloads on Kubernetes presents several challenges. Agent Substrate is a new project, built on and with Kubernetes, which enables you to securely run huge numbers of agents on Kubernetes with ultra low latency.
Kubernetes Solutions for Agent-Shaped Problems Nov 10, 2026 • 4:40 PM-5:10 PM Salt Palace | Level 1 | Hall DE
Agents and agent-like workloads are becoming a predominant class of workloads on Kubernetes, but they are different from "traditional" apps. In the last 18 months, those differences have been starkly emphasized. For over a decade K8s has been the industry standard platform for workloads of all sorts, with a rich ecosystem that has successfully adapted to every new class of workload. It’s time to adapt again. The Agent Substrate (http://ate.dev) project was born to tackle the problem of agent-like workloads head-on. Launched in May, Agent Substrate’s primary goal is to make it easy and efficient to run agent-like workloads. It leverages Kubernetes for what it is best at, then builds up from there to address the particular needs of agents. A lightweight control-plane coupled with active session management and network-triggered wake-ups enables users to run orders of magnitude more agent-like workloads than before – more efficiently and with lower latency and higher throughput.

Jeff Hoffman

OCI VP of Container Services

Oracle

Jeff Hoffman is a Vice President at Oracle, where he leads the Container, Kubernetes, and Serverless organization for Oracle Cloud Infrastructure (OCI). His team is responsible for OCI services including Oracle Kubernetes Engine (OKE), Container Instances, and Functions. Jeff has over 30 years of experience in cloud and engineering leadership, including work prior to Oracle at Amazon, Samsung, HashiCorp, and Dell.

LinkedIn

Conference Sessions

Production AI on Kubernetes: Built to Scale with Oracle Cloud Nov 11, 2026 • 10:04 AM-10:09 AM Salt Palace | Level 1 | Hall DE
Kubernetes is the foundation of the next generation of AI. Moving from experimentation to a production requires infrastructure designed for scale, performance and simplicity. In this keynote, Oracle Cloud Infrastructure (OCI) will explore how OCI Kubernetes Engine (OKE) enables organization to build and run demanding cloud-native and AI/ML workloads on an open, enterprise-ready platform. Join Jeff Hoffman, VP of Container Services at Oracle, to hear how customers are using OKE to bring AI initiatives into production. Learn how OCI’s Kubernetes platform, high-performance infrastructure, and open-source ecosystem help teams innovate faster while maintaining the control and reliability their critical workloads require.

Shaun Hopper

Production Engineer

Conference Sessions

Kubernetes with Conviction: How Customization is Shaping the OSS AI Ecosystem's Future Nov 12, 2026 • 9:56 AM-10:01 AM Salt Palace | Level 1 | Hall DE
Just as the Linux ecosystem thrives by separating a universal, low-level kernel from specialized distributions, Kubernetes is evolving beyond a monolithic, "one-size-fits-all" architecture. Forcing a single platform to act as a universal orchestrator for everything from tiny edge devices to hyper-scale AI clusters creates inefficiency and strain. To solve this, the industry is implicitly defining a lean Kubernetes Kernel—the core API machinery, reconciliation loops, and resource abstractions—as a standardized foundation. Simultaneously, agentic engineering is drastically reducing the cost of software creation and democratizing the operational expertise needed to build upon this foundation. Join us to learn how building on the Kubernetes Kernel allows power-users to create vertically integrated, purpose-built custom distros that achieve significant performance gains while still crowd-sourcing defect discovery and reliability.

Andjelko Iharos

VP of Architecture

HAProxy Technologies

Andjelko is the VP of Architecture at HAProxy Technologies. He is responsible for designing and overseeing the implementation of solutions and services at all scales—from low-level, microsecond-optimized software to massive automated clusters processing tens of billions of requests per day. Andjelko works with tools such as Python, C, Golang and many more.

Conference Sessions

Sovereign AI Starts at the Traffic Layer Nov 11, 2026 • 10:16 AM-10:21 AM Salt Palace | Level 1 | Hall DE
Cloud native teams spent a decade building traffic infrastructure they control: open source, decentralized, observable, and running on Kubernetes. Then AI arrived. Inference, the most sensitive layer in the stack, became a call to a vendor API. Your cluster is sovereign. Your AI isn't. You already run what's needed to reclaim that layer. An inference request is plain HTTP, with the model name in the JSON body. The ingress and Gateway API patterns in the cluster today can route and balance that traffic, and apply policy at the boundary. One data plane, instead of a fragmented traffic stack. This session covers consistent-hashing across GPU backends and model-aware routing from request inspection, with inference staying inside the cluster. Everything runs on HAProxy, the open source, model-agnostic load balancer and ingress controller built for high throughput and low latency. You keep sovereign control of AI traffic with no performance penalty.

Hao Jiang

Software Engineer

Hark

Hao Jiang is a Software Engineer at Hark and an open-source contributor to Ray and KubeRay. He focuses on building scalable infrastructure for multi-modal model training on Kubernetes. Previously, he worked on large-scale distributed storage and database systems at xAI, Mooncake Labs, and Nuro. He is also a PMC member of Apache OpenDAL and a leading contributor to the DuckDB ecosystem.

Conference Sessions

Scaling Multi-Modal Models and Computer-Use Agents on Kubernetes Nov 12, 2026 • 9:32 AM-9:37 AM Salt Palace | Level 1 | Hall DE
Hark builds multi-modal models and computer-use agents — models that operate software the way a person does. Scaling them means running every kind of workload on the same infrastructure: multi-modal data processing, training, reinforcement learning, and inference serving. All of it runs on Ray, deployed and autoscaled by KubeRay. But Ray is not only our runtime — it is also our resource management layer, with MultiKueue governing GPU capacity across every cluster we operate. This keynote covers how that fits together.

Ron Kahn

Senior Software Engineer

NVIDIA

Ron Kahn is a Senior Software Engineer at NVIDIA, a maintainer of NVIDIA Grove project, and snapshot project. When not simplifying AI for Kubernetes, Ron can be found cooking something delicious.

Conference Sessions

Kubeflow Graduation: Powering the Next Era of Cloud Native AI With an Open Community Nov 12, 2026 • 9:12 AM-9:17 AM Salt Palace | Level 1 | Hall DE
Since 2017, Kubeflow has evolved from a TensorFlow controller into a Cloud Native AI platform comprising eleven full-fledged projects, used by thousands of organizations across the AI lifecycle. Today, we are excited to celebrate Kubeflow’s Graduation from the CNCF – and the open community that made this journey possible. This keynote explores real-world stories from our users, showing how an open, vendor-neutral community is pushing the boundaries of Data & AI infrastructure across the whole AI lifecycle. It will highlight large-scale distributed training with Kubeflow Trainer and MPI, and look ahead to the agentic era with innovations such as the Semantic Operator. By bringing governed semantic APIs to Kubernetes, the Semantic Operator enables AI agents, applications, and BI tools to work with trusted business definitions without relying on LLM-generated SQL. Together, these efforts illustrate how the Kubeflow community is building the foundations for the next decade of Cloud Native AI.

Ekin Karabulut

Developer Advocate

NVIDIA

Ekin is a Developer Advocate at NVIDIA, following the acquisition of Run:ai. Prior to that, she specialized in the privacy implications of federated learning systems with DNNs. Currently, she is exploring the efficient usage of large models in production and one of the maintainers of KAI scheduler, which is a CNCF sandbox project.

Conference Sessions

End-User Panel – KubeFlow Enterprise Use Cases-Solutions Nov 11, 2026 • 2:08 PM-2:28 PM Salt Palace | Level 1 | Solutions Showcase | Halls 5 | Cosmic Canyon Zone | Cloud Native Theater
Panelists will discuss how their organizations leverage Kubeflow to orchestrate AI workloads on Kubernetes—from distributed training to low-latency serving—while overcoming enterprise challenges such as scalability, pipeline automation and hybrid-cloud integration. Attendees will gain actionable insights into architecture design, lessons learned from scaling AI workloads on Kubernetes, and proven patterns for accelerating model deployment from experimentation to production.
Kubeflow Graduation: Powering the Next Era of Cloud Native AI With an Open Community Nov 12, 2026 • 9:12 AM-9:17 AM Salt Palace | Level 1 | Hall DE
Since 2017, Kubeflow has evolved from a TensorFlow controller into a Cloud Native AI platform comprising eleven full-fledged projects, used by thousands of organizations across the AI lifecycle. Today, we are excited to celebrate Kubeflow’s Graduation from the CNCF – and the open community that made this journey possible. This keynote explores real-world stories from our users, showing how an open, vendor-neutral community is pushing the boundaries of Data & AI infrastructure across the whole AI lifecycle. It will highlight large-scale distributed training with Kubeflow Trainer and MPI, and look ahead to the agentic era with innovations such as the Semantic Operator. By bringing governed semantic APIs to Kubernetes, the Semantic Operator enables AI agents, applications, and BI tools to work with trusted business definitions without relying on LLM-generated SQL. Together, these efforts illustrate how the Kubeflow community is building the foundations for the next decade of Cloud Native AI.

Viv Keswani

Senior Director, Engineering

Uber

Viv Keswani is currently leading the Machine Learning/AI platform at Uber, known as Michelangelo, as well as the Search, Storage, and Data platforms. Before joining Uber, he spearheaded engineering teams across multiple domains, including ML/AI, Cloud, Storage, Networking, and Security, at both large corporations and startups. Viv holds a Master’s degree from Carnegie Mellon University and a Bachelor’s degree from the University of Mumbai, India.

Conference Sessions

Built on Kubernetes. Proven at Uber Scale. Now Open Source. Nov 11, 2026 • 10:11 AM-10:14 AM Salt Palace | Level 1 | Hall DE
Uber announces the open sourcing of Michelangelo AI, its Kubernetes-native ML platform designed to run across on-premises and public cloud environments, powering mission-critical AI across its global business. Battle-tested for nearly a decade, Michelangelo powers everything from traditional ML models to advanced generative AI and agentic workflows, supporting over 5,000 production models and 45 million predictions per second. Now, we're opening up the platform behind Uber's AI innovation to developers everywhere, with the goal of building an ecosystem where we can all innovate faster, together. Join us as we unveil Michelangelo AI and take a look behind the scenes at the platform powering AI at Uber's scale.

Ravin Kumar

Staff Research Engineer

Google Deepmind

Ravin is a Research Engineer at Google, with a long history in the hacker community, as a responsible member of his local hacker space attending DEFCON and other security conferences for the last 10 years. He was one of the leads of the AI GRT events and continues to work at the intersection of AI and security.

Conference Sessions

Pods in the Wild: Reinforcement Learning on Kubernetes with OpenRL Nov 10, 2026 • 9:32 AM-9:40 AM Salt Palace | Level 1 | Hall DE
What do wild dolphins, open-weight Gemma models, and Kubernetes have in common? Pods, of course! By training DolphinGemma on decades of marine acoustic data, researchers are actively decoding dolphin communication. This reflects a broader industry shift: transformative AI pairs domain-specific open weights with specialized data. Far from replacing frontier models, winning architectures unite massive foundation LLMs for broad reasoning with lightweight, task-tuned models for targeted execution. While frontier models set the bar for general intelligence, the next major operational leap belongs to everyday builders: empowering practitioners to fine-tune open-weight models that outperform massive frontier baselines on domain-specific tasks at a fraction of the cost, footprint, and latency. Historically, reinforcement learning (RL) was trapped inside elite research labs and brittle, complex runtimes. A new CNCF sandbox project, OpenRL fundamentally changes this paradigm by democratizing RL directly on Kubernetes. Now, a builder can articulate a task, point to a dataset, and use CNCF technologies to guide an open-weight model toward task mastery without requiring deep machine learning expertise.

Idit Levine

CEO

Solo.io

Idit Levine is the founder and CEO of Solo.io, where she is dedicated to simplifying agentic AI and empowering organizations to embrace cloud-native technologies. With a strong background in distributed systems and cloud infrastructure, Idit is passionate about transforming complex challenges into accessible, practical solutions for teams of all sizes by leveraging the power of innovative open-source technologies.

Conference Sessions

The Harness Is the New Runtime Nov 10, 2026 • 10:12 AM-10:17 AM Salt Palace | Level 1 | Hall DE
Agent harnesses are the dominant pattern for agentic AI on the desktop and will become the dominant runtime for agents on Kubernetes. The hard part is the move: bringing the harness onto your cluster with end user experience intact while adopting Kubernetes primitives for sandboxed execution, checkpoint/restore, scale, and security. Containers made this exact trip from developer machines to shared infrastructure in Kubernetes. In this keynote we'll present harnessed agents as your cluster's next critical workload. We'll show how harnesses combine models with a modular runtime of tools, skills, and plugins, and how leading CNCF ecosystem projects deliver the identity, policy, and observability to run them in production.

Frank Li

Head of Developer Relations

StepFun

Frank Li is Head of Global Developer Relations at StepFun, where he works with inference platforms, AI agent builders, developer tools, and open-source projects to bring StepFun models into real-world use. He has experience across multiple startups, and contributed to open-source data infrastructure.

Conference Sessions

The Storage Bottleneck AI Didn't See Coming: Inside StepFun’s Multimodal Production Stack Nov 12, 2026 • 9:51 AM-9:54 AM Salt Palace | Level 1 | Hall DE
Multimodal AI has outgrown the storage systems built for it. Training and inference now mix small-file storms, high-concurrency random reads, terabyte-scale checkpoints, and 100 GB+ model files across heterogeneous GPU clusters, private data centers, and public clouds. The hard part isn't any one workload. It means serving all of them through a single storage access layer that follows compute wherever it runs, whether on-prem or in any cloud. In this keynote, we'll show how StepFun made that move on Kubernetes, TiKV, and JuiceFS. Hardware-aware CSI mounts handle heterogeneous nodes, TiKV-backed metadata scales to more than 10,000 mounting hosts, point-get optimization speeds up relevant read-only operations by roughly 50%, and tuning both the metadata engine and readahead for Safetensors loads cuts read amplification by roughly 71%. We'll share how these improvements were merged upstream to turn production lessons into community gains.

Jago Macleod

Engineering Director, Kubernetes & GKE

Google

Jago Macleod is an Engineering Director at Google, where he leads much of the Kubernetes and Google Kubernetes Engine (GKE) team, which gives him the opportunity to work with some of Google Cloud’s largest customers. Prior to working at Google, Jago helped make the smart homes that may eventually turn against us at Nest; and before that, he made some of the games that will distract us along the way as VP of Engineering at Disney Interactive Games.

Conference Sessions

Kubernetes with Conviction: How Customization is Shaping the OSS AI Ecosystem's Future Nov 12, 2026 • 9:56 AM-10:01 AM Salt Palace | Level 1 | Hall DE
Just as the Linux ecosystem thrives by separating a universal, low-level kernel from specialized distributions, Kubernetes is evolving beyond a monolithic, "one-size-fits-all" architecture. Forcing a single platform to act as a universal orchestrator for everything from tiny edge devices to hyper-scale AI clusters creates inefficiency and strain. To solve this, the industry is implicitly defining a lean Kubernetes Kernel—the core API machinery, reconciliation loops, and resource abstractions—as a standardized foundation. Simultaneously, agentic engineering is drastically reducing the cost of software creation and democratizing the operational expertise needed to build upon this foundation. Join us to learn how building on the Kubernetes Kernel allows power-users to create vertically integrated, purpose-built custom distros that achieve significant performance gains while still crowd-sourcing defect discovery and reliability.

Manabu McCloskey

Sr. Open Source Engineer

AWS

Manabu is a Solutions Architect at Amazon Web Services. He focuses on contributing to open source application delivery tooling and works with AWS strategic customers to design and implement enterprise solutions using AWS resources and open source technologies. His interests include Kubernetes, GitOps, Serverless, and Souls Series.

LinkedIn

Conference Sessions

Stop AI Agents from Guessing SQL with Semantic Operator on Kubernetes Nov 11, 2026 • 1:58 PM-2:03 PM Salt Palace | Level 1 | Solutions Showcase | Halls 5 | Cosmic Canyon Zone | Cloud Native Theater
If you are building AI agents or BI applications that help analysts and business leaders answer questions from enterprise databases, we would value your feedback on a new open source project called Semantic Operator. AI agents can generate plausible SQL while misunderstanding business metrics, choosing incorrect joins, or bypassing governance rules. Semantic Operator addresses this by running Apache Ossie semantic models as Kubernetes resources. It provides agents and applications with certified metrics and dimensions, then generates deterministic, governed SQL for StarRocks or Trino. In this live demo, we will show how to manage semantic models through a Kubernetes operator, validate schemas and detect physical drift, enforce row, column, and metric governance, plan deterministic queries, and expose MCP and REST access for AI agents and applications. We will also walk through the end to end architecture and potential integration points.
Kubeflow Graduation: Powering the Next Era of Cloud Native AI With an Open Community Nov 12, 2026 • 9:12 AM-9:17 AM Salt Palace | Level 1 | Hall DE
Since 2017, Kubeflow has evolved from a TensorFlow controller into a Cloud Native AI platform comprising eleven full-fledged projects, used by thousands of organizations across the AI lifecycle. Today, we are excited to celebrate Kubeflow’s Graduation from the CNCF – and the open community that made this journey possible. This keynote explores real-world stories from our users, showing how an open, vendor-neutral community is pushing the boundaries of Data & AI infrastructure across the whole AI lifecycle. It will highlight large-scale distributed training with Kubeflow Trainer and MPI, and look ahead to the agentic era with innovations such as the Semantic Operator. By bringing governed semantic APIs to Kubernetes, the Semantic Operator enables AI agents, applications, and BI tools to work with trusted business definitions without relying on LLM-generated SQL. Together, these efforts illustrate how the Kubeflow community is building the foundations for the next decade of Cloud Native AI.

Phill Morton

Platform Architect

The Access Group

Phill is a dedicated software engineer turned platform engineer, with a strong passion for automating processes and enabling team success. With extensive experience in app modernization, cloud engineering, performance tuning, and observability. Phill brings a wealth of knowledge to the table. Throughout their career, Phill has been committed to demystifying complex processes, making them more accessible for team members, and fostering an environment where innovation and collaboration thrive.

Conference Sessions

The Flywheel Effect: Why Community Is the Ultimate End-User Advantage Nov 11, 2026 • 9:00 AM-9:10 AM Salt Palace | Level 1 | Hall DE
Cloud native promises a future without limits, but every end-user organization eventually discovers the same truth: no team becomes limitless alone. Every day, end users push CNCF projects into new industries, at new scales, and with entirely new challenges. They uncover new patterns, expose new edge cases, and find better ways to build. By sharing those experiences back with the community through talks, bug reports, reference architectures, and open collaboration, they create a flywheel where every contribution accelerates the next innovation. The result is a community where shared experience helps everyone solve problems faster, see around corners, and innovate with greater confidence. Together, Phill Morton and Abby Bangser will explore how end-user experience and community leadership fuel that flywheel. Through the upcoming Platform Engineering Community White Paper and a live demonstration of an enterprise AI platform built on those same principles, they’ll show how community ideas translate into real-world impact, enabling thousands of non-engineers to build at the speed of agents. You’ll leave with practical ideas for applying those patterns in your own organization and a new perspective on what “Cloud Native Without Limits” really means: the greatest advantage in cloud native isn’t just the technology we build; it’s the community we build it with.

Phong Nguyen

Production Engineer

Meta Platforms

Phong is a Production Engineer at Meta, bridging Meta’s production stack with public cloud and the wider cloud-native world. He ships platforms for everything from standard services to GPU workloads, then grabs the pager when things catch fire. And yes – when the smoke clears, he’s yapping about what worked so you don’t repeat the pain.

Conference Sessions

We Have GitOps at Home: How Meta Went from Terraform on Cron to GitOps for AI Infra Nov 11, 2026 • 9:57 AM-10:02 AM Salt Palace | Level 1 | Hall DE
At Meta, ~15 engineers run ~100 Kubernetes clusters for AI research & inference. Terraform exceeds expectations at its job: setting up infra for those clusters. The problem is its side gig: deploying the k8s apps. Even a one-line Helm values change can take a working day in our monolithic Terraform pipeline, where one stuck app can block dozens of others. In the new model, Terraform didn't get fired, just reorged. Kargo and Argo CD take over the apps, so each app ships on its own and a routine update is live and healthy in minutes. Terraform's side gig now comes down to installing ArgoCD. That's the easy part. The hard part is making GitOps work with Meta's internal systems: one-way config syncs from our in-house VCS to Git, an app "order form" instead of hand-written YAMLs, and multi-tenant ArgoCD(s) instead of one per cluster. If your company also runs on many homegrown systems, our lessons can save you a few headaches.

Shiva Nimmagadda

Vice President, Software Engineering

Salesforce

Engineering executive with 20+ years of experience building and operating large-scale cloud and AI platforms, leading engineering teams, and serving as Vice President, Engineering at Salesforce

LinkedIn

Conference Sessions

Salesforce Runs on Heroku: How We Empower Every Builder at Cloud Native Scale Nov 12, 2026 • 9:46 AM-9:49 AM Salt Palace | Level 1 | Hall DE
Salesforce is Heroku's largest customer with over 20k apps. Internal teams rely on Heroku to ship fast . As AI and "vibe coding" lower the barrier to software creation, it's essential to provide the right guardrails and restricted access while preserving an intuitive developer experience. To satisfy strict security requirements, Salesforce relies on Heroku Private Spaces. Behind the scenes, this isolated environment was re-engineered on a Cloud Native stack while keeping the same developer experience. Cloud Native Buildpacks bring the language & framework of choice, while image rebasing lets the platform patch base OS vulnerabilities across thousands of apps in seconds without rebuilding code. In this keynote, see how these pieces come together to provide a clear shared responsibility between the platform and the builder: builders own delivering customer value, while the platform silently handles scale, isolation, and security.

Robert Nishihara

Co-founder

Anyscale

Robert Nishihara is one of the creators of Ray, the world’s leading open source AI compute engine. Ray is used by companies across the board from Canva to Runway to Coinbase to Pinterest to scale their machine learning training, inference, and data processing workloads. He is one of the co-founders of Anyscale, which is the company behind Ray. He did his PhD in machine learning and distributed systems in the computer science department at UC Berkeley. Before that, he majored in math at Harvard.

Conference Sessions

Welcome + Opening Remarks Nov 10, 2026 • 9:00 AM-9:30 AM Salt Palace | Level 1 | Hall DE
The Scheduling Stack: Real-World Lessons from Anthropic, Anyscale, Google, and CERN Nov 12, 2026 • 11:50 AM-12:20 PM Salt Palace | Level 1 | Grand Ballroom H
For years, Kubernetes treated individual pods as scheduling units, complicating the execution of complex distributed systems. With the explosive growth of AI workloads, the ecosystem demands a smarter control plane. In this panel, scheduling experts from Google, Anthropic, Anyscale and CERN will unpack the paradigm shift of Workload Aware Scheduling. By teaching core Kubernetes to natively understand higher-level application semantics, we are bridging the critical gap between the scheduler and higher-level frameworks. This upstream evolution is essential for optimizing resource sharing and performance in massive multi-tenant environments. Our panelists will discuss how these foundational blocks enable a cohesive multi-layer scheduling stack. Furthermore, we will explore how this crucial first step unlocks the next frontiers of orchestration including seamless disruption management and tighter integration with infrastructure autoscaling.

Jorge Palma

Principal PM Lead, Azure Kubernetes Service

Microsoft

Jorge is the Principal PM Lead for AKS where he serves thousands of customers and mission critical application and helped lead the service to become the fastest growing service in Azure’s history. Formerly he was the Technical Lead for App Dev and DevOps with the Azure Global Black Belt team and has implemented and architected solutions in Azure since 2012, spanning across different roles from Services, Tech Evangelism and Tech Sales. He holds a MSc in Computer Engineering.

Conference Sessions

Agentic AI, Assembled in the Open Nov 12, 2026 • 10:10 AM-10:15 AM Salt Palace | Level 1 | Hall DE
Agentic AI is moving from experiments to production, and Kubernetes is a natural place to think about how these workloads are run. Let's look at how open source can provide a common platform layer for agentic systems, allowing frameworks, orchestrators, compute substrates, and identity infrastructure to come together behind declarative APIs. The goal is a more consistent way to deploy, schedule, secure, and observe AI agents using the cloud native model teams already know.

Vince Prignano

Staff Engineer

Apple

Vince Prignano is a Staff Engineer at Apple, where he leads Kubernetes infrastructure spanning clouds and bare metal.

Conference Sessions

Code, Containers, and Clusters: The Next Chapter of Apple's Container Framework Nov 11, 2026 • 9:38 AM-9:48 AM Salt Palace | Level 1 | Hall DE
Run Linux containers on macOS. Build with Apple silicon. Deploy at scale. Last year at KubeCon, we introduced Apple’s open source container framework. We kept building. Container Machine: The newest addition to the framework brings new functionality to Apple silicon. Container machines provide a highly integrated Linux environment that works seamlessly on your Mac. It’s fast, lightweight, and persistent. Kubernetes Plugin: We’re previewing a plugin for running Kubernetes locally on Apple silicon. Provision high-fidelity local clusters using upstream-native tools, such as kubeadm, kind images, and Cilium. At Apple, containers run at massive scale, while local testing maintains reliability. Watch us code a Kubernetes controller, then deploy it to a local cluster spun up with the plugin. All on Apple silicon. Leave with the roadmap, hands-on experience, and an invitation to join the community.
BoF | Let’s Discuss: The Future of Apple Containers with Cloud Native Tooling Nov 11, 2026 • 11:50 AM-12:20 PM Salt Palace | Level 2 | 254
Apple's open source Containerization framework lets developers run Linux containers on macOS in lightweight virtual machines. With new support for Container Machines and an experimental plug-in for local Kubernetes clusters, the project brings native, secure, and fast cloud tooling to developers with a Mac. Join this session to meet some of the development team, share how you use containers today, offer feedback on the project’s roadmap, and discuss best practices for using it as part of the cloud native ecosystem.

Luiz Scheidegger

Member of Technical Staff

Anthropic

Luiz is an engineer at Anthropic on the Claude Managed Agents team, building the sandboxing infrastructure that runs Claude agents. Before that he was Head of Engineering at Lindy.ai and an engineer at Nextdoor, where he built Launch Control, the company’s experimentation platform. Earlier he worked at Facebook and Uber.

Conference Sessions

Agent Substrate: Open Layers for Agentic Systems Nov 10, 2026 • 10:03 AM-10:10 AM Salt Palace | Level 1 | Hall DE
Running agentic workloads on Kubernetes presents several challenges. Agent Substrate is a new project, built on and with Kubernetes, which enables you to securely run huge numbers of agents on Kubernetes with ultra low latency.

Brandon Schoen

Senior Director, Professional Services

Heroku

Brandon is a Senior Director at Heroku, leading Professional Services and guiding product strategy. Dedicated to customer success, he helps teams design, build, and run applications on a fully managed, cloud-native platform. Brandon collaborates closely with engineering and client teams to empower both professional developers and citizen builders to reach production securely and at scale. Outside of work, he can be found in Las Vegas, NV, plotting his next grill upgrade.

LinkedIn

Conference Sessions

Salesforce Runs on Heroku: How We Empower Every Builder at Cloud Native Scale Nov 12, 2026 • 9:46 AM-9:49 AM Salt Palace | Level 1 | Hall DE
Salesforce is Heroku's largest customer with over 20k apps. Internal teams rely on Heroku to ship fast . As AI and "vibe coding" lower the barrier to software creation, it's essential to provide the right guardrails and restricted access while preserving an intuitive developer experience. To satisfy strict security requirements, Salesforce relies on Heroku Private Spaces. Behind the scenes, this isolated environment was re-engineered on a Cloud Native stack while keeping the same developer experience. Cloud Native Buildpacks bring the language & framework of choice, while image rebasing lets the platform patch base OS vulnerabilities across thousands of apps in seconds without rebuilding code. In this keynote, see how these pieces come together to provide a clear shared responsibility between the platform and the builder: builders own delivering customer value, while the platform silently handles scale, isolation, and security.

Abdel Sghiouar

Senior Cloud Developer Advocate

Google Cloud

Abdel Sghiouar is a senior Cloud Developer Advocate @Google Cloud. A co-host of the Kubernetes Podcast by Google and a CNCF Ambassador. His focused areas are GKE/Kubernetes, Service Mesh and Serverless. Abdel started his career in datacenters and infrastructure in Morocco where he is originally.

Conference Sessions

The Identity Crisis: Who is the AI Agent in Your Cluster? Nov 10, 2026 • 10:26 AM-10:36 AM Salt Palace | Level 1 | Hall DE
Kubernetes was designed for a predictable world. We built it to manage deterministic, stateless microservices with known communication paths, defined resource requirements, and static identity boundaries. The rise of agentic AI completely shatters these assumptions. An AI agent is not just another stateless workload, it is a non-deterministic entity that generates dynamic network calls, triggers external APIs, and executes machine-generated code on the fly. This behavior introduces a security blind spot: when an autonomous agent invokes a local tool or accesses a cloud database, what identity does it present? In this keynote, Abdel Sghiouar unpacks the looming identity and resource crisis of running agents on Kubernetes. He will explore why traditional Authn/Authz and container boundaries fail when software behaves unpredictably, and showcase how emerging open-source standards are trying to redefine multi-tenancy and secure execution for the next decade of cloud-native computing.
Your Voice on Stage: The Delivery + Storytelling Lab Nov 12, 2026 • 11:00 AM-12:00 PM Salt Palace | Level 1 | Grand Ballroom J
Getting your proposal accepted is only step one towards public speaking, now comes the real challenge: stepping onto the stage. Whether in front of 50 people or 5,000, delivering a memorable technical talk requires more than good slides; it demands storytelling, presence, and confidence. Join seasoned speakers Abdel Sghiouar and Julia Furst Morgado for a practical, hands-on delivery workshop. Moving beyond proposal writing, this session focuses on the art of stagecraft. You’ll learn how to structure technical concepts into compelling narrative arcs, use vocal range and body language to command the room, and apply proven strategies to manage stage fright and imposter syndrome. Through micro-drills, peer practice, and live "hot seat" coaching, you’ll test delivery techniques in real time. Walk away with actionable frameworks, increase confidence, and the stagecraft skills needed to make your next presentation unforgettable!

Yodar Shafrir

CEO

ScaleOps

Conference Sessions

Context Unlocks Autonomous AI Infrastructure Nov 11, 2026 • 9:50 AM-9:55 AM Salt Palace | Level 1 | Hall DE
Enterprise AI never sits still. Thousands of agents, tools, models, and applications spike, idle, scale, and redeploy every minute, each with its own needs, each affecting the others. Managing production AI infrastructure manually has become impossible. The answer is infrastructure that runs autonomously, and that depends on real-time context: knowing how each workload behaves, what it needs, and what it talks to before any action is taken. Yodar Shafrir shows how context, from a single workload to the whole system, becomes autonomous orchestration that keeps enterprise AI reliable and performant, and puts platform teams in command of the system instead of chasing it.

Jeffrey Sica

Head of Projects

CNCF

Jeffrey (@jeefy) is Head of Projects at the CNCF, with a focus on improving community and project automation. Before that, he worked at Red Hat and the University of Michigan focusing on Cloud Native technologies and CI/CD patterns. Jeffrey has been a contributor to upstream Kubernetes, helping in SIG-Contribex, SIG-Release, and SIG-UI. He passionately advocates for open source development and recognizing and alleviating burnout.

Conference Sessions

Cloud Native: Built by Humans, Engineered to Evolve Nov 12, 2026 • 9:00 AM-9:10 AM Salt Palace | Level 1 | Hall DE
Cloud native has never stood still. As workloads, infrastructure and the way we build software change, an open ecosystem of projects and people gives us the ability to evolve with them. Jeffrey Sica will give a look at how the breadth of the cloud native landscape and the strength of its community are helping us adapt to AI, increasingly specialized infrastructure and whatever comes next.

Mike Stefaniak

Senior Manager of Product

AWS

Mike Stefaniak is a Senior Manager of Product at AWS, where he leads product strategy for containers. His team focuses on simplifying Kubernetes for platform engineers and application developers at scale, spanning compute elasticity, workload abstraction, and cluster governance.

Conference Sessions

Engineering Kubernetes for the Next Era of Workloads Nov 10, 2026 • 9:42 AM-9:47 AM Salt Palace | Level 1 | Hall DE
AI changed what Kubernetes has to handle: container images in the tens of gigabytes, clusters approaching 100,000 nodes, and accelerators a device count can’t describe. At AWS we hit those limits early, and when we do, we fix them upstream. This keynote covers three layers. In containerd, parallel layer unpack takes the long pole out of pulling large AI images. In the control plane, we removed lock contention in the API server watch cache and the Horizontal Pod Autoscaler and fixed a scheduler hot path, all in Kubernetes 1.37. For accelerators, Dynamic Resource Allocation lets a pod ask for a GPU and network adapter on the same PCIe root, and Karpenter can now provision for it, even for nodes that don’t exist yet.

Lin Sun

Head of Open Source

Solo.io

Lin is the Head of Open Source at Solo.io, contributing full-time to the open-source community. She serves on the CNCF Technical Oversight Committee (TOC), is a CNCF Ambassador, and is a maintainer for Istio, kgateway, and kagent. An international speaker at tech conferences, Lin frequently blogs about gateways, service meshes, cloud-native connectivity, and AI agents. She is the author of “Sidecar-less Istio Explained” and “AI Agents in Kubernetes”, and holds more than 200 patents.

Conference Sessions

Cloud Native Without Limits: What Can AI Agents Do for Us? Nov 12, 2026 • 10:17 AM-10:27 AM Salt Palace | Level 1 | Hall DE
In Amsterdam, we explored what cloud native could do with AI agents, MCP, and a drone controlled through natural language. Now it's time for Drone 2. What happens when AI agents become a natural part of the cloud native ecosystem? Imagine agents that can debug production issues, analyze token spending, automate routine operations, and even make us laugh, all while interacting with Kubernetes, cloud native tools, and MCP servers through natural language, with humans firmly in control. This keynote explores what's possible when AI agents become trusted collaborators in cloud native. Through live demonstrations with Drone 2, we'll show how natural language and AI agents can transform the way we build, operate, and experience cloud native systems, pushing the boundaries of what's possible while keeping humans in control.
Agentic + Cloud Native: Getting Started Contributing to Kagent Nov 12, 2026 • 11:00 AM-12:20 PM Salt Palace | Level 2 | 255 A
Curious how kagent brings agentic AI to the cloud native ecosystem? Interested in contributing to kagent ? Want to explore the codebase and understand what powers the kagent framework? Join the kagent maintainers for a hands-on contributor session where you'll learn how the project is built and how you can become part of its growing open source community. We'll walk through the architecture of kagent, explore how agents and MCP tools work together, show you how to set up a local development environment, explain the project's community and contribution workflow, and help you get started with your first contribution. Whether you're new to the project or looking to become a regular contributor, you'll leave with the knowledge and confidence to make your first PR to kagent.

Lucy Sweet

Member of Technical Staff

Anthropic

Lucy is a Member of Technical Staff at Anthropic and co-lead of the Kubernetes Node Lifecycle Working Group.

Conference Sessions

A Tribute to Open Source Nov 11, 2026 • 9:31 AM-9:36 AM Salt Palace | Level 1 | Hall DE
Kubernetes The Hard Way – Live Nov 11, 2026 • 2:20 PM-2:50 PM Salt Palace | Level 1 | Grand Ballroom GI
Come behind the scenes of how Kubernetes actually works! In this session, Lucy Sweet (Member of Technical Staff at Anthropic and Node Lifecycle WG co-Lead for the Kubernetes project) and Sandeep Kanabar (Lead Software Engineer at Gen) will take you on a live journey through setting up Kubernetes “The Hard Way”. No tools, no automation. We’ll peel back the layers of abstraction to reveal how the core components of Kubernetes interact to orchestrate workloads from small to planet scale. We’ll trace the lifecycle of a Pod from creation through to running container, and explain how often-overlooked components play a critical role in cluster operations. Through this live demonstration, attendees will gain a deep understanding of Kubernetes' internal mechanics, empowering them to troubleshoot more effectively, make informed architectural decisions, and unlock the true potential of Kubernetes in production environments.
The Future of Kubernetes Eviction and Node Lifecycle Nov 11, 2026 • 5:00 PM-5:30 PM Salt Palace | Level 2 | 251 A-C
Kubernetes disruption handling is evolving from a collection of custom solutions and workarounds into a more cooperative lifecycle model. Join the Node Lifecycle Working Group for an update on two key efforts: the Eviction Request API for workload disruption coordination, and Specialized Lifecycle Management. We will discuss how the new declarative Eviction API gives workloads, controllers, and administrators a more explicit way to negotiate disruption than today’s best-effort eviction paths. We will explore how it fits alongside PDBs, API-initiated eviction, preemption, autoscaling, and how it can help with app migration and maintenance automation. We will also cover Node Lifecycle Conditions, the v1alpha foundation for Specialized Lifecycle Management. This change introduces new conditions to Nodes for draining and node maintenance. A shared condition model gives Kubernetes components and ecosystem controllers a common place to observe the Node lifecycle.

Andrey Velichkevich

AI Engineer

Apple

Andrey Velichkevich is an AI Engineer at Apple and a key contributor to the Kubeflow and Kubernetes projects. He serves on the Kubeflow Steering Committee and is a co-chair for Kubernetes WG WAS and Checkpoint/Restore. In addition, Andrey is an active member of the Batch WG, SIG Scheduling, and the CNCF AI TSG. He is a co-author of the CNCF AI White Paper and actively contributes to multiple AI initiatives that advance the CNCF community and the broader cloud-native AI ecosystem.

Conference Sessions

Kubeflow Summit Opening Remarks Nov 11, 2026 • 12:45 PM-12:50 PM Salt Palace | Level 1 | Solutions Showcase | Halls 5 | Cosmic Canyon Zone | Cloud Native Theater
A Decade of AI on Kubernetes: Kubeflow’s Journey to CNCF Graduation Nov 11, 2026 • 3:10 PM-3:40 PM Salt Palace | Level 2 | 250 A-C
Over the past decade, Kubeflow has helped define how AI and machine learning run on Kubernetes, evolving from data engineering, distributed training, and model serving to supporting today's generative AI and agentic workloads. That journey has reached a major milestone with CNCF Graduation. In this session, Kubeflow maintainers will reflect on the project's evolution, the architectural and community decisions that shaped it, and the lessons learned in building a sustainable open source platform for enterprise-grade AI. The speakers will discuss how Kubeflow adapted to each wave of AI innovation while remaining Kubernetes-native, what CNCF Graduation means for governance, sustainability, and enterprise adoption, and what's next for the project. Attendees will gain a retrospective of the project's evolution, an overview of its current state, and insights into the future of AI on Kubernetes.
One Workload to Rule Them All: Workload Aware Scheduling for Distributed AI Training on Kubernetes Nov 11, 2026 • 5:00 PM-5:30 PM Salt Palace | Level 1 | Grand Ballroom ACE
Distributed AI training fails fast: if a single GPU pod in a multi-node job cannot schedule, accelerators sit idle or the workload fails with collective-op timeouts. To compensate, the ecosystem evolved many external schedulers, including Kueue, Volcano, and KAI, each forcing controllers to pick a side. Kubernetes 1.36 changes this. The new Workload and PodGroup APIs (KEP-4671) introduce scheduler-native primitives for gang-scheduling, topology-aware placement, and coordinated use of high-bandwidth interconnects such as IMEX. TrainJob, JobSet, and LeaderWorkerSet are converging on these APIs. This session covers Kubeflow Trainer’s adoption (KEP-3015): why TrainJob owns Workload composition and how its plugin architecture enables advanced scheduling features. Attendees will see a demo of multi-node fine-tuning on modern accelerators, how the Workload API unifies a fragmented ecosystem, and how to adopt it in their own controllers.
Kubeflow Graduation: Powering the Next Era of Cloud Native AI With an Open Community Nov 12, 2026 • 9:12 AM-9:17 AM Salt Palace | Level 1 | Hall DE
Since 2017, Kubeflow has evolved from a TensorFlow controller into a Cloud Native AI platform comprising eleven full-fledged projects, used by thousands of organizations across the AI lifecycle. Today, we are excited to celebrate Kubeflow’s Graduation from the CNCF – and the open community that made this journey possible. This keynote explores real-world stories from our users, showing how an open, vendor-neutral community is pushing the boundaries of Data & AI infrastructure across the whole AI lifecycle. It will highlight large-scale distributed training with Kubeflow Trainer and MPI, and look ahead to the agentic era with innovations such as the Semantic Operator. By bringing governed semantic APIs to Kubernetes, the Semantic Operator enables AI agents, applications, and BI tools to work with trusted business definitions without relying on LLM-generated SQL. Together, these efforts illustrate how the Kubeflow community is building the foundations for the next decade of Cloud Native AI.
WG-Workload Aware Scheduling: An Introduction Nov 12, 2026 • 11:00 AM-11:30 AM Salt Palace | Level 2 | 250 A-C
The Kubernetes community has been hard at work enhancing the scheduling of workloads. In 2026, the community started adding support for workload aware scheduling and this work was eventually folded into a working group. In this session, we will highlight the APIs added, provide details on the in-tree roadmap and the goal of the working group with respect to the ecosystem. We will discuss the integrations with Kueue and popular workload controllers like JobSet. This session will provide the roadmap and future plans for the working group. We are also looking for engagement from the community.

Kevin Wang

Technical Expert, Lead of CloudNative Open Source

Huawei

Kevin Wang has been an outstanding contributor in the CNCF community since its beginning and is the leader of the cloud native open source team at Huawei. Kevin has contributed critical enhancements to Kubernetes, led the incubation of the KubeEdge, Volcano, Karmada projects in CNCF, actively participated in the TOC, and played a critical role in building and supporting the growth of CNCF technologies and the CNCF community in the globe.

Conference Sessions

Volcano for the Agentic AI Era: Unified Scheduling Across Training, Inference, and Agents Nov 10, 2026 • 3:40 PM-4:10 PM Salt Palace | Level 2 | 251 D-F
Agentic AI is changing what cloud-native infrastructure must schedule. AI training, inference, and RL workloads continue to stress scheduler throughput and coordination, while agent workloads add pressure on startup latency and burst handling. The question is no longer how to optimize one scheduling path, but how to balance throughput, latency, and resource utilization across shared Kubernetes clusters. In this maintainer session, we will share how Volcano is evolving from a batch-oriented scheduler into a unified scheduling framework for AI training, inference, emerging RL workloads, and agents. The focus is the new framework: a coordinated architecture that improves scheduler parallelism while handling conflicts among scheduler workers and keeping scheduling behavior consistent across workload paths. We will also cover community updates: benchmark results, topology-aware placement, Volcano Global, colocation improvements, and roadmap.
The Cloud Native Feedback Loop: How End Users and Developers Drive CNCF Projects Forward Nov 11, 2026 • 9:19 AM-9:29 AM Salt Palace | Level 1 | Hall DE
How do today’s reference architectures become real production platforms built on CNCF projects that evolve, mature, and earn long-term trust? In this joint keynote the CNCF End User Technical Advisory Board (TAB) and Technical Oversight Committee (TOC) – with representation from the TAB User Groups and TOC Technical Advisory Groups – cover the mapping reference architectures and real-world production needs to the project lifecycle decisions that shape the CNCF ecosystem. Along the way they connect this technical and governance journey to the data and tools that support smarter decisions at scale by showcasing User Group and Technical Advisory Group Initiatives, as well as LFX and LF Insights metrics.

Marlow Warnicke

Senior Software Engineer

NVIDIA

Marlow is a Principal Cloud Engineer working on scheduling at SchedMD. She also is a chair for the CNCF Environmental Sustainability TAG. Marlow has expertise in resource management, the AI/ML Kubernetes cloud compute ecosystem, embedded systems, high performance compute system tools, kernel drivers, tracing libraries, and security. Marlow's interests lie in optimizing the cloud native ecosystem for both performance and sustainability.

Conference Sessions

The Cloud Native Feedback Loop: How End Users and Developers Drive CNCF Projects Forward Nov 11, 2026 • 9:19 AM-9:29 AM Salt Palace | Level 1 | Hall DE
How do today’s reference architectures become real production platforms built on CNCF projects that evolve, mature, and earn long-term trust? In this joint keynote the CNCF End User Technical Advisory Board (TAB) and Technical Oversight Committee (TOC) – with representation from the TAB User Groups and TOC Technical Advisory Groups – cover the mapping reference architectures and real-world production needs to the project lifecycle decisions that shape the CNCF ecosystem. Along the way they connect this technical and governance journey to the data and tools that support smarter decisions at scale by showcasing User Group and Technical Advisory Group Initiatives, as well as LFX and LF Insights metrics.
Explore TAG Workloads Foundation: Core Runtime, Batch Scheduling, and Moar Nov 12, 2026 • 1:50 PM-2:20 PM Salt Palace | Level 2 | 251 A-C
TAG Workloads Foundation defines practices and standards for how cloud native systems execute and manage workloads: from container runtimes and schedulers all the way through to lifecycle management at scale. Come learn about projects tackling GPU sharing, VM-based workloads, batch workloads, and secure compute, and meet the people behind them. You'll also hear about ongoing work to document how to run scientific AI on Kubernetes at scale. Whether you run scientific computing, manage GPU clusters, or build infrastructure for AI teams, you'll leave with a clear picture of what cloud native workload execution actually looks like across the full stack and where to plug in.

Eitan Yarmush

Software Architect

Solo.io

Devoted network engineer and doggy dad. I spend my time hiking, painting, problem solving, and taking care of my new pup. Trying to make the infra space just a little better 🙂

Conference Sessions

Agent Substrate: Open Layers for Agentic Systems Nov 10, 2026 • 10:03 AM-10:10 AM Salt Palace | Level 1 | Hall DE
Running agentic workloads on Kubernetes presents several challenges. Agent Substrate is a new project, built on and with Kubernetes, which enables you to securely run huge numbers of agents on Kubernetes with ultra low latency.
Lessons Learned Sandboxing Agents with kagent Nov 10, 2026 • 5:30 PM-6:00 PM Salt Palace | Level 1 | Hall DE
Is an agent a normal Kubernetes workload? It needs compute, identity, networking, storage, and lifecycle management – so it looks like one. But an agent session is long-lived, stateful, bursty, and mostly idle. And it can do anything a user can. That's what makes sandboxing critical. The sandbox defines what the agent can see, touch, call, and persist: filesystem access, tool permissions, credentials, network policy, identity, and cleanup. Getting that boundary wrong is a security incident waiting to happen. This talk covers the lessons we learned designing and implementing agent sandboxing for kagent – including NVIDIA's OpenShell, Kubernetes' agent-sandbox, and Google's substrate support – and what teams need to think about before trusting agents with code execution, credentials, network access, and production data.
Agentic + Cloud Native: Getting Started Contributing to Kagent Nov 12, 2026 • 11:00 AM-12:20 PM Salt Palace | Level 2 | 255 A
Curious how kagent brings agentic AI to the cloud native ecosystem? Interested in contributing to kagent ? Want to explore the codebase and understand what powers the kagent framework? Join the kagent maintainers for a hands-on contributor session where you'll learn how the project is built and how you can become part of its growing open source community. We'll walk through the architecture of kagent, explore how agents and MCP tools work together, show you how to set up a local development environment, explain the project's community and contribution workflow, and help you get started with your first contribution. Whether you're new to the project or looking to become a regular contributor, you'll leave with the knowledge and confidence to make your first PR to kagent.

Joe Zhou

Lead Developer Advocate

Juicedata Inc.

Joe is a cloud-native advocate for storage systems and former startup CTO who excels at leading and motivating team members to deliver exceptional results. He currently serves as the Lead Developer Advocate in North America at JuiceFS (a high-performance distributed file system) and helps AI/ML engineers, developers, and platform teams get the most out of JuiceFS for performance and efficiency.

Conference Sessions

The Storage Bottleneck AI Didn't See Coming: Inside StepFun’s Multimodal Production Stack Nov 12, 2026 • 9:51 AM-9:54 AM Salt Palace | Level 1 | Hall DE
Multimodal AI has outgrown the storage systems built for it. Training and inference now mix small-file storms, high-concurrency random reads, terabyte-scale checkpoints, and 100 GB+ model files across heterogeneous GPU clusters, private data centers, and public clouds. The hard part isn't any one workload. It means serving all of them through a single storage access layer that follows compute wherever it runs, whether on-prem or in any cloud. In this keynote, we'll show how StepFun made that move on Kubernetes, TiKV, and JuiceFS. Hardware-aware CSI mounts handle heterogeneous nodes, TiKV-backed metadata scales to more than 10,000 mounting hosts, point-get optimization speeds up relevant read-only operations by roughly 50%, and tuning both the metadata engine and readahead for Safetensors loads cuts read amplification by roughly 71%. We'll share how these improvements were merged upstream to turn production lessons into community gains.