diff --git a/tests/engine/test_arg_utils.py b/tests/engine/test_arg_utils.py
index 10077d24a..8dec9c5a9 100644
--- a/tests/engine/test_arg_utils.py
+++ b/tests/engine/test_arg_utils.py
@@ -10,6 +10,7 @@ from types import SimpleNamespace
 from unittest.mock import Mock
 
 import pytest
+from omegaconf import OmegaConf
 from pydantic import ValidationError
 from transformers import PretrainedConfig
 from vllm.engine.arg_utils import EngineArgs
@@ -17,6 +18,7 @@ from vllm.engine.arg_utils import EngineArgs
 from vllm_omni.config.model import OmniModelConfig
 from vllm_omni.engine.arg_utils import OmniEngineArgs
 from vllm_omni.engine.async_omni_engine import AsyncOmniEngine
+from vllm_omni.engine.stage_init_utils import build_engine_args_dict
 
 pytestmark = [pytest.mark.core_model, pytest.mark.cpu]
 
@@ -314,3 +316,15 @@ def test_strip_single_engine_args_model_does_not_trigger_warning(mocker):
     assert "compilation_config" in warned_args
     assert "tensor_parallel_size" not in warned_args
     assert "model" not in warned_args
+
+
+# For https://github.com/vllm-project/vllm-omni/issues/3293
+def test_tensor_parallel_size_none_is_handled():
+    """Ensure the tensor parallel size of None isn't forwarded."""
+    engine_args = OmegaConf.create({"stage_id": 0, "engine_args": {"tensor_parallel_size": None}})
+    args = build_engine_args_dict(
+        engine_args,
+        model="snu-aidas/Dynin-Omni",
+    )
+    assert isinstance(args, dict)
+    assert "tensor_parallel_size" not in args
diff --git a/vllm_omni/engine/stage_init_utils.py b/vllm_omni/engine/stage_init_utils.py
index eb25f1577..277edb5eb 100644
--- a/vllm_omni/engine/stage_init_utils.py
+++ b/vllm_omni/engine/stage_init_utils.py
@@ -397,6 +397,13 @@ def build_engine_args_dict(
 ) -> dict[str, Any]:
     """Build the normalized engine args dict for one stage."""
     engine_args = stage_config.engine_args
+    # HACK (Alex) Tensor parallel size should not be passed as None;
+    # remove it if this is the case so that we fall back to default
+    # creation from vLLM's engine args.
+    # NOTE: This will be fixed more generically in ongoing work for engine arg filtering.
+    if "tensor_parallel_size" in engine_args and engine_args["tensor_parallel_size"] is None:
+        del engine_args["tensor_parallel_size"]
+
     stage_type = getattr(stage_config, "stage_type", "llm")
     stage_id = stage_config.stage_id