diff --git a/.buildkite/run-neuron-test.sh b/.buildkite/run-neuron-test.sh
index 9259391aaed49..aa29c434e7cfb 100644
--- a/.buildkite/run-neuron-test.sh
+++ b/.buildkite/run-neuron-test.sh
@@ -3,6 +3,18 @@
 # This script build the Neuron docker image and run the API server inside the container.
 # It serves a sanity check for compilation and basic model usage.
 set -e
+set -v
+
+image_name="neuron/vllm-ci"
+container_name="neuron_$(tr -dc A-Za-z0-9 < /dev/urandom | head -c 10; echo)"
+
+HF_CACHE="$(realpath ~)/huggingface"
+mkdir -p "${HF_CACHE}"
+HF_MOUNT="/root/.cache/huggingface"
+
+NEURON_COMPILE_CACHE_URL="$(realpath ~)/neuron_compile_cache"
+mkdir -p "${NEURON_COMPILE_CACHE_URL}"
+NEURON_COMPILE_CACHE_MOUNT="/root/.cache/neuron_compile_cache"
 
 # Try building the docker image
 aws ecr get-login-password --region us-west-2 | docker login --username AWS --password-stdin 763104351884.dkr.ecr.us-west-2.amazonaws.com
@@ -13,41 +25,30 @@ if [ -f /tmp/neuron-docker-build-timestamp ]; then
     last_build=$(cat /tmp/neuron-docker-build-timestamp)
     current_time=$(date +%s)
     if [ $((current_time - last_build)) -gt 86400 ]; then
+        docker image prune -f
         docker system prune -f
+        rm -rf "${HF_MOUNT:?}/*"
+        rm -rf "${NEURON_COMPILE_CACHE_MOUNT:?}/*"
         echo "$current_time" > /tmp/neuron-docker-build-timestamp
     fi
 else
     date "+%s" > /tmp/neuron-docker-build-timestamp
 fi
 
-docker build -t neuron -f Dockerfile.neuron .
+docker build -t "${image_name}" -f Dockerfile.neuron .
 
 # Setup cleanup
-remove_docker_container() { docker rm -f neuron || true; }
+remove_docker_container() {
+    docker image rm -f "${image_name}" || true;
+}
 trap remove_docker_container EXIT
-remove_docker_container
 
 # Run the image
-docker run --device=/dev/neuron0 --device=/dev/neuron1 --network host --name neuron neuron python3 -m vllm.entrypoints.api_server \
-       --model TinyLlama/TinyLlama-1.1B-Chat-v1.0 --max-num-seqs 8 --max-model-len 128 --block-size 128 --device neuron --tensor-parallel-size 2 &
-
-# Wait for the server to start
-wait_for_server_to_start() {
-    timeout=300
-    counter=0
-
-    while [ "$(curl -s -o /dev/null -w '%{http_code}' localhost:8000/health)" != "200" ]; do
-        sleep 1
-        counter=$((counter + 1))
-        if [ $counter -ge $timeout ]; then
-            echo "Timeout after $timeout seconds"
-            break
-        fi
-    done
-}
-wait_for_server_to_start
-
-# Test a simple prompt
-curl -X POST -H "Content-Type: application/json" \
-    localhost:8000/generate \
-    -d '{"prompt": "San Francisco is a"}'
+docker run --rm -it --device=/dev/neuron0 --device=/dev/neuron1 --network host \
+       -v "${HF_CACHE}:${HF_MOUNT}" \
+       -e "HF_HOME=${HF_MOUNT}" \
+       -v "${NEURON_COMPILE_CACHE_URL}:${NEURON_COMPILE_CACHE_MOUNT}" \
+       -e "NEURON_COMPILE_CACHE_URL=${NEURON_COMPILE_CACHE_MOUNT}" \
+       --name "${container_name}" \
+       ${image_name} \
+       /bin/bash -c "python3 /workspace/vllm/examples/offline_inference_neuron.py"
diff --git a/Dockerfile.neuron b/Dockerfile.neuron
index 269139fe90f0b..e9cb82889decd 100644
--- a/Dockerfile.neuron
+++ b/Dockerfile.neuron
@@ -15,8 +15,8 @@ RUN apt-get update && \
         ffmpeg libsm6 libxext6 libgl1
 
 ### Mount Point ###
-# When launching the container, mount the code directory to /app
-ARG APP_MOUNT=/app
+# When launching the container, mount the code directory to /workspace
+ARG APP_MOUNT=/workspace
 VOLUME [ ${APP_MOUNT} ]
 WORKDIR ${APP_MOUNT}/vllm
 
@@ -25,6 +25,7 @@ RUN python3 -m pip install --no-cache-dir fastapi ninja tokenizers pandas
 RUN python3 -m pip install sentencepiece transformers==4.45.2 -U
 RUN python3 -m pip install transformers-neuronx --extra-index-url=https://pip.repos.neuron.amazonaws.com -U
 RUN python3 -m pip install neuronx-cc==2.16.345.0 --extra-index-url=https://pip.repos.neuron.amazonaws.com -U
+RUN python3 -m pip install pytest
 
 COPY . .
 ARG GIT_REPO_CHECK=0
@@ -42,4 +43,7 @@ RUN --mount=type=bind,source=.git,target=.git \
 # install development dependencies (for testing)
 RUN python3 -m pip install -e tests/vllm_test_utils
 
+# overwrite entrypoint to run bash script
+RUN echo "import subprocess; import sys; subprocess.check_call(sys.argv[1:])" > /usr/local/bin/dockerd-entrypoint.py
+
 CMD ["/bin/bash"]
diff --git a/examples/offline_inference_neuron.py b/examples/offline_inference_neuron.py
index 2856be7c864ea..f098c8e5fed1e 100644
--- a/examples/offline_inference_neuron.py
+++ b/examples/offline_inference_neuron.py
@@ -1,12 +1,5 @@
-import os
-
 from vllm import LLM, SamplingParams
 
-# creates XLA hlo graphs for all the context length buckets.
-os.environ['NEURON_CONTEXT_LENGTH_BUCKETS'] = "128,512,1024,2048"
-# creates XLA hlo graphs for all the token gen buckets.
-os.environ['NEURON_TOKEN_GEN_BUCKETS'] = "128,512,1024,2048"
-
 # Sample prompts.
 prompts = [
     "Hello, my name is",
@@ -26,8 +19,8 @@
     # Currently, this is a known limitation in continuous batching support
     # in transformers-neuronx.
     # TODO(liangfu): Support paged-attention in transformers-neuronx.
-    max_model_len=2048,
-    block_size=2048,
+    max_model_len=1024,
+    block_size=1024,
     # The device can be automatically detected when AWS Neuron SDK is installed.
     # The device argument can be either unspecified for automated detection,
     # or explicitly assigned.