Apply remaining UX improvements from old branch

- Update backend READMEs with correct KVBM status
- Simplify Kubernetes README with cleaner structure
- Fix multinode example to use correct dynamo.sglang command
- Add missing --skip-tokenizer-init flags

Signed-off-by: athreesh <[email protected]>

components/backends/sglang/README.md

@@ -40,7 +40,7 @@ git checkout $(git describe --tags $(git rev-list --tags --max-count=1))
 | [**KV-Aware Routing**](../../../docs/architecture/kv_cache_routing.md) | ✅ |  |
 | [**SLA-Based Planner**](../../../docs/architecture/sla_planner.md) | ✅ |  |
 | [**Load Based Planner**](../../../docs/architecture/load_planner.md) | ❌ | Planned |
-| [**KVBM**](../../../docs/architecture/kvbm_architecture.md) | ❌ | Planned |
+| [**KVBM**](../../../docs/architecture/kvbm_architecture.md) | 🚧 | WIP |
 
 ### Large Scale P/D and WideEP Features
 

components/backends/trtllm/README.md

@@ -57,7 +57,7 @@ git checkout $(git describe --tags $(git rev-list --tags --max-count=1))
 | [**KV-Aware Routing**](../../../docs/architecture/kv_cache_routing.md) | ✅ |  |
 | [**SLA-Based Planner**](../../../docs/architecture/sla_planner.md) | ✅ |  |
 | [**Load Based Planner**](../../../docs/architecture/load_planner.md) | 🚧 | Planned |
-| [**KVBM**](../../../docs/architecture/kvbm_architecture.md) | 🚧 | Planned |
+| [**KVBM**](../../../docs/architecture/kvbm_architecture.md) | ✅ |  |
 
 ### Large Scale P/D and WideEP Features
 

components/backends/vllm/README.md

@@ -40,7 +40,7 @@ git checkout $(git describe --tags $(git rev-list --tags --max-count=1))
 | [**KV-Aware Routing**](../../../docs/architecture/kv_cache_routing.md) | ✅ |  |
 | [**SLA-Based Planner**](../../../docs/architecture/sla_planner.md) | ✅ |  |
 | [**Load Based Planner**](../../../docs/architecture/load_planner.md) | 🚧 | WIP |
-| [**KVBM**](../../../docs/architecture/kvbm_architecture.md) | 🚧 | WIP |
+| [**KVBM**](../../../docs/architecture/kvbm_architecture.md) | ✅ |  |
 | [**LMCache**](./LMCache_Integration.md) | ✅ |  |
 
 ### Large Scale P/D and WideEP Features

docs/kubernetes/README.md

@@ -24,7 +24,7 @@ High-level guide to Dynamo Kubernetes deployments. Start here, then dive into sp
 ```bash
 # 1. Set environment
 export NAMESPACE=dynamo-kubernetes
-export RELEASE_VERSION=0.x.x # any version of Dynamo 0.3.2+ listed at https://github.com/ai-dynamo/dynamo/releases
+export RELEASE_VERSION=0.5.0 # any version of Dynamo 0.3.2+ listed at https://github.com/ai-dynamo/dynamo/releases
 
 # 2. Install CRDs
 helm fetch https://helm.ngc.nvidia.com/nvidia/ai-dynamo/charts/dynamo-crds-${RELEASE_VERSION}.tgz
@@ -33,19 +33,51 @@ helm install dynamo-crds dynamo-crds-${RELEASE_VERSION}.tgz --namespace default
 # 3. Install Platform
 helm fetch https://helm.ngc.nvidia.com/nvidia/ai-dynamo/charts/dynamo-platform-${RELEASE_VERSION}.tgz
 helm install dynamo-platform dynamo-platform-${RELEASE_VERSION}.tgz --namespace ${NAMESPACE} --create-namespace
+
+# For multinode deployments, enable Grove/KAI with:
+# helm install dynamo-platform dynamo-platform-${RELEASE_VERSION}.tgz --namespace ${NAMESPACE} --create-namespace \
+#   --set grove.enabled=true \
+#   --set kai.enabled=true
 ```
 
 For more details or customization options (including multinode deployments), see **[Installation Guide for Dynamo Kubernetes Platform](/docs/kubernetes/installation_guide.md)**.
 
-## 2. Choose Your Backend
-
-Each backend has deployment examples and configuration options:
-
-| Backend | Available Configurations |
-|---------|--------------------------|
-| **[vLLM](/components/backends/vllm/deploy/README.md)** | Aggregated, Aggregated + Router, Disaggregated, Disaggregated + Router, Disaggregated + Planner, Disaggregated Multi-node |
-| **[SGLang](/components/backends/sglang/deploy/README.md)** | Aggregated, Aggregated + Router, Disaggregated, Disaggregated + Planner, Disaggregated Multi-node |
-| **[TensorRT-LLM](/components/backends/trtllm/deploy/README.md)** | Aggregated, Aggregated + Router, Disaggregated, Disaggregated + Router, Disaggregated Multi-node |
+## 2. Choose Your Backend and Deployment Pattern
+
+### **Aggregated Serving**
+Prefill and decode phases run on the same worker - simplest deployment pattern.
+
+| Backend | Configuration | Deploy Command |
+|---------|---------------|----------------|
+| **vLLM** | [Aggregated](components/backends/vllm/deploy/agg.yaml) | `kubectl apply -f components/backends/vllm/deploy/agg.yaml -n ${NAMESPACE}` |
+| **vLLM** | [Aggregated + Router](components/backends/vllm/deploy/agg_router.yaml) | `kubectl apply -f components/backends/vllm/deploy/agg_router.yaml -n ${NAMESPACE}` |
+| **SGLang** | [Aggregated](components/backends/sglang/deploy/agg.yaml) | `kubectl apply -f components/backends/sglang/deploy/agg.yaml -n ${NAMESPACE}` |
+| **SGLang** | [Aggregated + Router](components/backends/sglang/deploy/agg_router.yaml) | `kubectl apply -f components/backends/sglang/deploy/agg_router.yaml -n ${NAMESPACE}` |
+| **TensorRT-LLM** | [Aggregated](components/backends/trtllm/deploy/agg.yaml) | `kubectl apply -f components/backends/trtllm/deploy/agg.yaml -n ${NAMESPACE}` |
+| **TensorRT-LLM** | [Aggregated + Router](components/backends/trtllm/deploy/agg_router.yaml) | `kubectl apply -f components/backends/trtllm/deploy/agg_router.yaml -n ${NAMESPACE}` |
+
+### **Disaggregated Serving**
+Prefill and decode phases run on separate workers - higher performance and scalability.
+
+| Backend | Configuration | Deploy Command |
+|---------|---------------|----------------|
+| **vLLM** | [Disaggregated](components/backends/vllm/deploy/disagg.yaml) | `kubectl apply -f components/backends/vllm/deploy/disagg.yaml -n ${NAMESPACE}` |
+| **vLLM** | [Disaggregated + Router](components/backends/vllm/deploy/disagg_router.yaml) | `kubectl apply -f components/backends/vllm/deploy/disagg_router.yaml -n ${NAMESPACE}` |
+| **vLLM** | [Disaggregated + Planner](components/backends/vllm/deploy/disagg_planner.yaml) | `kubectl apply -f components/backends/vllm/deploy/disagg_planner.yaml -n ${NAMESPACE}` |
+| **SGLang** | [Disaggregated](components/backends/sglang/deploy/disagg.yaml) | `kubectl apply -f components/backends/sglang/deploy/disagg.yaml -n ${NAMESPACE}` |
+| **SGLang** | [Disaggregated + Planner](components/backends/sglang/deploy/disagg_planner.yaml) | `kubectl apply -f components/backends/sglang/deploy/disagg_planner.yaml -n ${NAMESPACE}` |
+| **TensorRT-LLM** | [Disaggregated](components/backends/trtllm/deploy/disagg.yaml) | `kubectl apply -f components/backends/trtllm/deploy/disagg.yaml -n ${NAMESPACE}` |
+| **TensorRT-LLM** | [Disaggregated + Router](components/backends/trtllm/deploy/disagg_router.yaml) | `kubectl apply -f components/backends/trtllm/deploy/disagg_router.yaml -n ${NAMESPACE}` |
+| **TensorRT-LLM** | [Disaggregated + Planner](components/backends/trtllm/deploy/disagg_planner.yaml) | `kubectl apply -f components/backends/trtllm/deploy/disagg_planner.yaml -n ${NAMESPACE}` |
+
+### **Multi-node Deployment** (Model replicaes across multiple nodes)
+Scale disaggregated serving across multiple Kubernetes nodes for maximum performance.
+
+| Backend | Configuration | Deploy Command |
+|---------|---------------|----------------|
+| **vLLM** | [Multi-node](components/backends/vllm/deploy/disagg-multinode.yaml) | `kubectl apply -f components/backends/vllm/deploy/disagg-multinode.yaml -n ${NAMESPACE}` |
+| **SGLang** | [Multi-node](components/backends/sglang/deploy/disagg-multinode.yaml) | `kubectl apply -f components/backends/sglang/deploy/disagg-multinode.yaml -n ${NAMESPACE}` |
+| **TensorRT-LLM** | [Multi-node](components/backends/trtllm/deploy/disagg-multinode.yaml) | `kubectl apply -f components/backends/trtllm/deploy/disagg-multinode.yaml -n ${NAMESPACE}` |
 
 ## 3. Deploy Your First Model
 

examples/basics/multinode/README.md

@@ -131,7 +131,7 @@ Open a terminal on Node 1 and launch both workers:
 
 ```bash
 # Launch prefill worker in background
-CUDA_VISIBLE_DEVICES=0 python3 -m dynamo.sglang.worker \
+CUDA_VISIBLE_DEVICES=0 python3 -m dynamo.sglang \
     --model-path Qwen/Qwen3-0.6B \
     --served-model-name Qwen/Qwen3-0.6B \
     --page-size 16 \
@@ -141,7 +141,7 @@ CUDA_VISIBLE_DEVICES=0 python3 -m dynamo.sglang.worker \
     --disaggregation-mode prefill \
     --disaggregation-transfer-backend nixl &
 
-CUDA_VISIBLE_DEVICES=1 python3 -m dynamo.sglang.decode_worker \
+CUDA_VISIBLE_DEVICES=1 python3 -m dynamo.sglang \
     --model-path Qwen/Qwen3-0.6B \
     --served-model-name Qwen/Qwen3-0.6B \
     --page-size 16 \
@@ -157,15 +157,15 @@ CUDA_VISIBLE_DEVICES=1 python3 -m dynamo.sglang.decode_worker \
 > - `--page-size 16`: Sets the KV cache block size - must be identical across all workers
 > - `--disaggregation-mode`: Separates prefill (prompt processing) from decode (token > generation)
 > - `--disaggregation-transfer-backend nixl`: Enables high-speed GPU-to-GPU transfers
-> - `--skip-tokenizer-init`: Avoids duplicate tokenizer loading since the frontend > handles tokenization
+> - `--skip-tokenizer-init`: Avoids duplicate tokenizer loading since the frontend handles tokenization
 
 ### Step 3: Launch Replica 2 (Node 2)
 
 Open a terminal on Node 2 and launch both workers:
 
 ```bash
 # Launch prefill worker in background
-CUDA_VISIBLE_DEVICES=0 python3 -m dynamo.sglang.worker \
+CUDA_VISIBLE_DEVICES=0 python3 -m dynamo.sglang \
     --model-path Qwen/Qwen3-0.6B \
     --served-model-name Qwen/Qwen3-0.6B \
     --page-size 16 \
@@ -176,7 +176,7 @@ CUDA_VISIBLE_DEVICES=0 python3 -m dynamo.sglang.worker \
     --disaggregation-transfer-backend nixl &
 
 # Launch decode worker in foreground
-CUDA_VISIBLE_DEVICES=1 python3 -m dynamo.sglang.decode_worker \
+CUDA_VISIBLE_DEVICES=1 python3 -m dynamo.sglang \
     --model-path Qwen/Qwen3-0.6B \
     --served-model-name Qwen/Qwen3-0.6B \
     --page-size 16 \
@@ -473,7 +473,7 @@ Stop all components in reverse order:
      exit
 
      # Method 3: Kill by process name (from any terminal)
-     pkill -f "dynamo.sglang.worker.*prefill"
+     pkill -f "dynamo.sglang.*prefill"
      ```
 3. Stop infrastructure services:
    ```bash