diff --git a/torchtitan/experiments/compiler_toolkit/common_utils.py b/torchtitan/experiments/compiler_toolkit/common_utils.py
deleted file mode 100644
index df997aabe9..0000000000
--- a/torchtitan/experiments/compiler_toolkit/common_utils.py
+++ /dev/null
@@ -1,20 +0,0 @@
-# Copyright (c) Meta Platforms, Inc. and affiliates.
-# All rights reserved.
-#
-# This source code is licensed under the BSD-style license found in the
-# LICENSE file in the root directory of this source tree.
-
-from contextlib import contextmanager
-
-from torchtitan.config import JobConfig
-
-
-@contextmanager
-def disable_compile(job_config: JobConfig):
-    """Context manager to temporarily disable compilation."""
-    original_value = job_config.compile.enable
-    job_config.compile.enable = False
-    try:
-        yield
-    finally:
-        job_config.compile.enable = original_value
diff --git a/torchtitan/experiments/compiler_toolkit/deepseek_v3/parallelize.py b/torchtitan/experiments/compiler_toolkit/deepseek_v3/parallelize.py
index a859415c1c..3e4dff5013 100644
--- a/torchtitan/experiments/compiler_toolkit/deepseek_v3/parallelize.py
+++ b/torchtitan/experiments/compiler_toolkit/deepseek_v3/parallelize.py
@@ -18,7 +18,6 @@
 from torchtitan.config import JobConfig
 from torchtitan.distributed import ParallelDims
 from torchtitan.distributed.expert_parallel import ExpertParallel
-from torchtitan.experiments.compiler_toolkit.common_utils import disable_compile
 
 from torchtitan.experiments.compiler_toolkit.graph_utils import (
     CompiledModule,
@@ -103,12 +102,13 @@ def parallelize_deepseekv3(
     parallel_dims: ParallelDims,
     job_config: JobConfig,
 ) -> CompiledModule:
+    assert (
+        not job_config.compile.enable
+    ), "compile.enable should be False in the compiler toolkit style workflow."
 
     annotate_model()
 
-    # Disable torch.compile over the model in the compiler toolkit style workflow
-    with disable_compile(job_config):
-        model = simple_fsdp_parallelize_deepseekv3(model, parallel_dims, job_config)
+    model = simple_fsdp_parallelize_deepseekv3(model, parallel_dims, job_config)
 
     # TODO: CompiledModule should take sample input as well, so that we can
     # compile ahead of time.
diff --git a/torchtitan/experiments/compiler_toolkit/llama3/parallelize.py b/torchtitan/experiments/compiler_toolkit/llama3/parallelize.py
index b8d00db39e..8ee937324c 100644
--- a/torchtitan/experiments/compiler_toolkit/llama3/parallelize.py
+++ b/torchtitan/experiments/compiler_toolkit/llama3/parallelize.py
@@ -15,7 +15,6 @@
 
 from torchtitan.config import JobConfig
 from torchtitan.distributed import ParallelDims
-from torchtitan.experiments.compiler_toolkit.common_utils import disable_compile
 
 from torchtitan.experiments.compiler_toolkit.graph_utils import (
     CompiledModule,
@@ -105,12 +104,13 @@ def parallelize_llama(
     parallel_dims: ParallelDims,
     job_config: JobConfig,
 ) -> CompiledModule:
+    assert (
+        not job_config.compile.enable
+    ), "compile.enable should be False in the compiler toolkit style workflow."
 
     annotate_model()
 
-    # Disable torch.compile over the model in the compiler toolkit style workflow
-    with disable_compile(job_config):
-        model = simple_fsdp_parallelize_llama(model, parallel_dims, job_config)
+    model = simple_fsdp_parallelize_llama(model, parallel_dims, job_config)
 
     # TODO: CompiledModule should take sample input as well, so that we can
     # compile ahead of time.