fix: self review fixes

Author: AlonKellner-Jounce

src/guidellm/scheduler/worker.py

@@ -195,15 +195,17 @@ async def resolve_scheduler_request(
             await asyncio.sleep(wait_time)
 
         info.worker_start = time.time()
-        request_start_result: WorkerProcessResult[RequestT, ResponseT] = (
-            WorkerProcessResult(
-                type_="request_start",
-                request=request,
-                response=None,
-                info=info,
+        asyncio.create_task(
+            self.send_result(
+                results_queue,
+                WorkerProcessResult(
+                    type_="request_start",
+                    request=request,
+                    response=None,
+                    info=info,
+                ),
             )
         )
-        asyncio.create_task(self.send_result(results_queue, request_start_result))
 
         status, response = await self.resolve(request, timeout_time)
         info.worker_end = time.time()

tests/e2e/vllm_sim_server.py

@@ -10,8 +10,8 @@
 
 class VllmSimServer:
     """
-    A class to manage the lifecycle of a test server.
-    Encapsulates starting, checking health, and stopping the server process.
+    [vLLM simulator](https://llm-d.ai/docs/architecture/Components/inf-simulator)
+    A vLLM simulator wrapper for pytest.
     """
 
     def __init__(