feat(landing): add Open Source Summit Korea event page

src/components/EventLanding.js

@@ -114,6 +114,7 @@ export default function EventLanding({ slug }) {
                   <FontAwesomeIcon icon={faClock} className={styles.metaIcon} />
                   {event.startTime}
                   {event.endTime ? ` - ${event.endTime}` : ""}
+                  {event.timeZone ? ` ${event.timeZone}` : ""}
                 </span>
               )}
               {event.room && (

src/data/events.js

@@ -139,6 +139,36 @@ const events = [
     externalUrl: "https://coscup.org/2026/en/",
     talkUrl: "https://coscup.org/2026/en/session/BGYZ3B",
   },
+  {
+    slug: "opensource-summit-korea",
+    title: {
+      en: "Simplifying AI for Edge Compute with HAMi",
+      zh: "使用 HAMi 简化边缘计算中的 AI",
+    },
+    date: "2026-08-12",
+    startTime: "11:40",
+    endTime: "12:10",
+    timeZone: "KST",
+    location: {
+      en: "Grand InterContinental Seoul Parnas",
+      zh: "首尔柏纳斯大洲际酒店",
+    },
+    address: {
+      addressLocality: "Seoul",
+      addressCountry: "KR",
+      streetAddress: "521 Teheran-ro Gangnam-gu, 06164",
+    },
+    description: {
+      en: "Open-source agents like Hermes and OpenClaw can reason and use tools, but edge deployment is still hard - not the model, the compute underneath: limited memory, tight power budgets, no ops team. Democratizing agentic AI means fixing the compute layer. This deep dive covers GPU and memory slicing at the edge: carving one device's unified memory so multiple agents run concurrently on hardware as small as an NVIDIA Jetson, and time-sharing when demand exceeds memory. We compare Jetson-class GPUs with NPUs from Axelera and DeepX - higher performance per watt - and what it takes to make them schedulable. HAMi, the CNCF Incubation project for hardware-agnostic GPU virtualization, is that layer: one scheduling plane across heterogeneous accelerators, fine-grained memory slicing, open source. A blueprint for edge AI without a cloud budget.",
+      zh: "开源智能体（如 Hermes 和 OpenClaw）能够推理并使用工具，但边缘部署仍然困难——问题不在模型，而在底层的算力：内存有限、功耗预算紧张、没有运维团队。让智能体 AI 普及化，关键在于修复算力层。本次深度解析涵盖边缘 GPU 与显存切片：将一块设备的统一内存切分，让多个智能体在 NVIDIA Jetson 这样的小型硬件上并发运行，并在内存不足时进行时间共享。我们对比 Jetson 级 GPU 与 Axelera、DeepX 的 NPU——每瓦性能更高——以及让它们可被调度的必要条件。HAMi 是 CNCF 孵化项目，提供与硬件无关的 GPU 虚拟化，正是这一层：跨异构加速器的统一调度平面、细粒度显存切片、开源。一份无需云预算的边缘 AI 蓝图。",
+    },
+    cta: {
+      discordUrl: "https://go.dynamia.ai/hami-chat-korea",
+      githubUrl: "https://go.dynamia.ai/proj-hami-korea",
+    },
+    externalUrl: "https://events.linuxfoundation.org/open-source-summit-korea/",
+    talkUrl: "https://events.linuxfoundation.org/open-source-summit-korea/program/schedule/",
+  },
 ];
 
 export default events;

src/pages/landing/opensource-summit-korea.js

@@ -0,0 +1,3 @@
+import EventLanding from "@site/src/components/EventLanding";
+
+export default () => <EventLanding slug="opensource-summit-korea" />;