Update video descriptions and enhance video grid layout (#129)

- Improved video descriptions to include presenter names and affiliations for better context.
- Refactored the video grid layout from a grid display to a flexbox layout for improved responsiveness and alignment.
- Added new styles for video cards to enhance visual presentation and user interaction.

Signed-off-by: Pete Cheslock <[email protected]>

Author: petecheslock

src/pages/videos.js

@@ -7,22 +7,27 @@ const videos = [
   {
     id: 'lvzMLf6wXlY',
     title: 'Kubernetes Native Distributed Inferencing',
-    description: 'Introduction to llm-d at DevConf.US 2025 — learn the fundamentals of distributed LLM inference on Kubernetes.',
+    description: 'Introduction to llm-d at DevConf.US 2025 — learn the fundamentals of distributed LLM inference on Kubernetes from Rob Shaw (Red Hat).',
   },
   {
     id: 'mfXIe_S53vA',
     title: 'Serving PyTorch LLMs at Scale',
-    description: 'Disaggregated inference with Kubernetes and llm-d — presented by M. Ayoub & C. Liu at PyTorch Conference.',
+    description: 'Disaggregated inference with Kubernetes and llm-d — presented by Maroon Ayoub (IBM) & Cong Liu (Google) at PyTorch Conference.',
   },
   {
     id: '_xAXb70d4-0',
     title: 'Distributed Inference with Well-Lit Paths',
-    description: 'Explore llm-d\'s "well-lit paths" approach to simplified, production-ready distributed inference.',
+    description: 'Watch Rob Shaw (Red Hat) explore llm-d\'s "well-lit paths" and its approach to simplified, production-ready distributed inference.',
   },
   {
     id: 'g8_snJA_ESU',
     title: 'Multi-Accelerator LLM Inference',
-    description: 'Deep dive into multi-accelerator LLM inference on Kubernetes — presented by Erwan Gallen at KubeCon.',
+    description: 'Deep dive into multi-accelerator LLM inference on Kubernetes — presented by Erwan Gallen (Red Hat) at KubeCon.',
+  },
+  {
+    id: '-C76naL3PRc',
+    title: 'Routing Stateful AI Workloads in Kubernetes',
+    description: 'Maroon Ayoub (IBM) & Michey Mehta (Red Hat) explore cache-aware routing strategies for LLM workloads using llm-d and the K8s Gateway API Inference Extension.',
   },
 ];
 

src/pages/videos.module.css

@@ -82,14 +82,32 @@
 
 /* Video Grid */
 .videoGrid {
-  display: grid;
-  grid-template-columns: repeat(2, 1fr);
+  display: flex;
+  flex-wrap: wrap;
+  justify-content: center;
   gap: 2.5rem;
 }
 
+/* Video Card */
+.videoCard {
+  flex: 0 1 calc(50% - 1.25rem);
+  max-width: calc(50% - 1.25rem);
+  background: #fff;
+  border-radius: 16px;
+  overflow: hidden;
+  box-shadow: 
+    0 4px 6px rgba(127, 49, 127, 0.06),
+    0 12px 24px rgba(127, 49, 127, 0.08);
+  transition: transform 0.3s ease, box-shadow 0.3s ease;
+}
+
 @media screen and (max-width: 900px) {
+  .videoCard {
+    flex: 0 1 100%;
+    max-width: 100%;
+  }
+  
   .videoGrid {
-    grid-template-columns: 1fr;
     gap: 2rem;
   }
 
@@ -102,17 +120,6 @@
   }
 }
 
-/* Video Card */
-.videoCard {
-  background: #fff;
-  border-radius: 16px;
-  overflow: hidden;
-  box-shadow: 
-    0 4px 6px rgba(127, 49, 127, 0.06),
-    0 12px 24px rgba(127, 49, 127, 0.08);
-  transition: transform 0.3s ease, box-shadow 0.3s ease;
-}
-
 .videoCard:hover {
   transform: translateY(-6px);
   box-shadow: