DaoCloud

Three core AI inference cluster challenges: GPU topology, scheduling, cost validation. 🎤DaoCloud's Weizhou Lan presents topology discovery for Kueue scheduling, using Kwok virtual nodes for cost-free validation to boost GPU and cluster efficiency. 📅March 26,2026 14:30-15:00CET 📍Hall 8 | Room D

Growing AI workloads such as LLMs make cache-aware scheduling and routing critical to GPU efficiency. 🎤Kay Yan (DaoCloud) collaborates with Red Hat & IBM to resolve K8s KV-Cache fragmentation, sharing high-performance LLM inference best practices. 📅March 26,2026 11:00-12:15CET 📍 Elicium 1

📢KubeEdge has become the preferred platform for Kubernetes edge scaling! 🎤 DaoCloud's Hongbing Zhang will collaborate with Huawei and VMware experts to share edge workload management architecture, multi-domain implementation cases, and bring the latest technical updates and community dynamics. 📅March 25,2026 11:45-12:15CET 📍 Amtrium 1+2

🌐 At the KubeCon China 2025 Opening Keynote, DaoCloud's CEO and Founder, Roby Chen 🤝 HKGAI Director Assistant, Head of Tech, Dr. Aaron Xu delivered speech on localized innovation in AI for Hong Kong 📈 Tracing HKGAI V1 → HKGAI Studio evolution ⚡ Revealing how open source reshape HK's AI ecosystem #KubeCon #AI #HKGAI #DaoCloud

🌟 On June 10, 2025, #KubeCon + #CloudNativeCon China 2025 opened at the Hopewell Hotel in Hong Kong. The program co-chairs, Roby Chen (Founder & CEO of #DaoCloud and #CNCF Ambassador), and Vivian Hu (Product Manager of Second State) welcomed thousands of developers and cloud native enthusiasts from across the globe. Jim Zemlin, Executive Director of the Linux Foundation, declared a milestone moment: "Open source AI has won". Open source is not only cutting infrastructure costs but also accelerating the evolution of AI agents and driving innovation at the application layer. Citing data from Harvard Business School, Jim emphasized that open source software delivers an astounding $9 trillion in global economic value. Over the past decade, with 285,000 developers across 190+ countries and regions, the CNCF community has made over 3,350,000+ code commits, 1,200,000+ pull requests, and 18,800000+ contributions, driving more than 200 open source projects from incubation to global adoption. Chris Aniszczyk, CTO of CNCF, highlighted that Chinese developers now rank 2nd among the top 15 contributor countries, with 1,068,553 contributions to the CNCF community. Five of the top ten Chinese contributors to Kubernetes hail from DaoCloud: Paco Xu, Kante Yin, Shiming Zhang, Michael Yao, and Baofa Fan. Beyond code, Chinese developers are leading the incubation of globally influential CNCF projects, such as #Volcano and #HAMi, both of which are pivotal to the advancement of AI workloads. To recognize and expand global collaboration, CNCF launched the Cloud Native Ambassador Program. Outstanding Chinese contributors like Roby Chen, Iceber Gu, and Michael Yao have already been inducted. The next round of applications is open through the end of June—developers passionate about shaping the future of cloud native are encouraged to apply. CNCF also introduced the Kubestronaut Program to honor individuals with deep technical expertise and exceptional contributions to cloud native education. China now boasts 29 Kubestronauts, including Iceber Gu from DaoCloud. In a joint keynote, Roby Chen and Aaron Xu, Head of Technology at #HKGAI, showcased the vision of #HKGAI. Roby emphasized that #Kubernetes has matured into the “big brother” of AI infrastructure, driving efficiency, security, and cost savings. Aaron outlined the city’s AI development strategy around four pillars: data, compute, model, and governance. Together, they unveiled HKGAI Studio, a full-stack AI infrastructure for Hong Kong, built on DaoCloud's cloud native technologies. Aaron announced that dozens of applications built on this platform will launch in July to serve millions of Hong Kong residents. DaoCloud developers draw crowds at booth S5 with live demos of the d.run computing power scheduling platform and DeepSeek All-in-one. Through real-world case studies, they show how cloud native technologies empower scalable and efficient AI infrastructure.

Hong Kong, June 9, 2025 – KubeCon + CloudNativeCon China 2025 will be held at the Hopewell Hotel in Hong Kong on June 10-11, bringing together industry leaders, project maintainers, and end users from around the globe. The program co-chairs, Roby Chen (Founder & CEO of DaoCloud, CNCF Ambassador) and Vivian Hu (from Second State), along with a global panel of experts, have meticulously curated the conference agenda. From a pool of 571 submissions, they selected the top 10%—cutting-edge sessions with significant industry value. These sessions will cover the latest advancements in cloud native and AI, practical implementations of model inference technologies, advanced Kubernetes cluster optimization, real-world applications of open source projects, and more. The conference will kick off on June 10th with an opening keynote by Roby Chen alongside Aaron Xu (Assistant to Director, Head of Technology, HKGAI). Using the launch of HKGAI V1 as a springboard, they will explore what lies ahead for AI development in Hong Kong. As a pioneer in leveraging cloud native technologies to advance AI adoption, Roby Chen will share insights on building the essential "highway" for AI infrastructure based on the open source ecosystem. Aaron Xu will contribute the perspective of Hong Kong's AI innovation practices, highlighting the region's unique advantages in fundamental AI research and cross-border collaboration. Twelve developers from DaoCloud will deliver fourteen sessions. Highlights include presentations on widely anticipated open source projects in LLM inference such as vLLM and LWS, as well as the innovative K8sGPT project, which empowers Kubernetes with LLM technologies. Paco Xu, a member of the Kubernetes Steering Committee, and community representatives will host a live "Newcomer Bootcamp," offering a hands-on, step-by-step tutorial on making community contributions – a must-attend session for newcomers to the community. DaoCloud also unveils Dataset, a practical and efficient tool meticulously crafted for the data preparation stage of model development. Additional exciting content will cover areas like observability, cloud edge collaboration, and chaos engineering. Stay tuned for the detailed agenda. Beyond the enriching sessions, attendees are invited to visit the S5 booth to explore optimal solutions for enterprise AI privatization. Experience firsthand the cutting-edge technologies behind the d.run computing power scheduling platform and the advanced capabilities of the AI appliance. These are designed to eliminate computing bottlenecks, streamline application deployment, and take the next step toward efficient, scalable, and controllable enterprise-grade AI.

Either in Jupyter notebook or LLM serving, python dependency is always a headache？ Three developers from DaoCloud will introduce Dataset, a practical and efficient tool meticulously crafted for the data preparation stage of LLM development. #KubeCon2025 #KubeCon #CloudNative #AI #LLM

As AI workloads like LLMs grow, scalable GPU clusters are essential！ Kay Yan from DaoCloud will share insights from deploying custom GPU clusters at scale with Kubespray, discussing challenges and best practices. #KubeCon2025 #KubeCon #CloudNative #AI #LLM

How did KubeEdge evolve from a small project into a diverse community？ ✨Hongbing Zhang will lead everyone to explore how to build a multi-vendor and diverse community, and how to expand into different industries, with other core maintainers. #KubeEdge #KubeCon2025

Are you aware of KubeEdge, the first Graduated project in edge computing？ 💡In this session, Hongbing Zhang from DaoCloud will present an overview of its architecture for extending Kubernetes to the edge and showcase its successful industry applications.#KubeEdge #Edge #KubeCon2025