diff --git a/cpp/tensorrt_llm/thop/moe/communication/moeAlltoAllOp.cpp b/cpp/tensorrt_llm/thop/moe/communication/moeAlltoAllOp.cpp
index b4be9d9aa6a9..93b17d123a34 100644
--- a/cpp/tensorrt_llm/thop/moe/communication/moeAlltoAllOp.cpp
+++ b/cpp/tensorrt_llm/thop/moe/communication/moeAlltoAllOp.cpp
@@ -74,7 +74,7 @@ static constexpr size_t CACHELINE_ALIGNMENT = 128;
 //
 // The MoE all-to-all workspace setup used tensorrt_llm::mpi::MpiComm directly.
 // mpiUtils.h throws "MPI is disabled, DON'T USE MPI" whenever TLLM_DISABLE_MPI=1,
-// which the Ray executor sets for every worker (executor/ray_executor.py), so the
+// which the Ray executor sets for every worker (tensorrt_llm/executor/ray/executor.py), so the
 // NVLinkOneSided strategy always failed there and silently degraded to
 // NVLinkTwoSided. Route through the Torch process group in that case; the same
 // pattern is already used by cacheTransceiver / allreduceOp via pg_utils.
diff --git a/tensorrt_llm/_torch/attention/backends/fp4_mla/cache_update.py b/tensorrt_llm/_torch/attention/backends/fp4_mla/cache_update.py
index 47c80e861731..de869807805d 100644
--- a/tensorrt_llm/_torch/attention/backends/fp4_mla/cache_update.py
+++ b/tensorrt_llm/_torch/attention/backends/fp4_mla/cache_update.py
@@ -787,9 +787,7 @@ def scatter_fp4_mla_kv_cache(
     and ``positions[token_offset : token_offset + N]``. Callers must pass a
     latent_cache pre-sliced to the current phase (context or generation) so
     that ``shape[0]`` matches the number of index entries they intend to
-    consume. ``MLA.forward_impl`` (tensorrt_llm/_torch/modules/attention.py)
-    slices ``latent_cache[:num_ctx_tokens]`` for context and
-    ``latent_cache[num_ctx_tokens:]`` for generation before dispatching.
+    consume.
 
     Callers must pass ``phase``, ``local_layer``, and ``v_head_dim``. Context
     scatter writes the final FP4 tile representation directly. Dimensions
diff --git a/tensorrt_llm/_torch/pyexecutor/kv_cache/mamba_cache_manager/mamba_cache_manager_v2.py b/tensorrt_llm/_torch/pyexecutor/kv_cache/mamba_cache_manager/mamba_cache_manager_v2.py
index 518bd08ca37b..8b47da347877 100644
--- a/tensorrt_llm/_torch/pyexecutor/kv_cache/mamba_cache_manager/mamba_cache_manager_v2.py
+++ b/tensorrt_llm/_torch/pyexecutor/kv_cache/mamba_cache_manager/mamba_cache_manager_v2.py
@@ -93,7 +93,7 @@
 )
 
 if TYPE_CHECKING:
-    from tensorrt_llm._torch.attention_backend.interface import AttentionMetadata
+    from tensorrt_llm._torch.attention.backends.interface import AttentionMetadata
     from tensorrt_llm.llmapi.llm_args import DecodingBaseConfig
     from tensorrt_llm.sampling_params import SamplingParams
 
diff --git a/tensorrt_llm/_torch/speculative/dflash.py b/tensorrt_llm/_torch/speculative/dflash.py
index b2bd10c8fb2b..5e57da4634b7 100644
--- a/tensorrt_llm/_torch/speculative/dflash.py
+++ b/tensorrt_llm/_torch/speculative/dflash.py
@@ -1277,7 +1277,8 @@ def _init_ctx_block_tables(
         ``kv_cache_block_offsets`` entries are ``pool_block_index *
         num_pool_layers * kv_factor`` plus the K/V field index. What that has to
         be divided by to index a ``get_buffers`` view depends on the manager --
-        see the stride check below; attention_backend/utils.py hardcodes the V1
+        see the stride check below;
+        ``append_mla_latent_cache_generation_cuda_graph_safe`` hardcodes the V1
         divisor and is not a general contract.
         """
         mgr = draft_kv_cache_manager
diff --git a/tests/unittest/_torch/ray_orchestrator/multi_gpu/test_mnnvl_allreduce.py b/tests/unittest/_torch/ray_orchestrator/multi_gpu/test_mnnvl_allreduce.py
index ac9da13a1dda..9222fce91c02 100644
--- a/tests/unittest/_torch/ray_orchestrator/multi_gpu/test_mnnvl_allreduce.py
+++ b/tests/unittest/_torch/ray_orchestrator/multi_gpu/test_mnnvl_allreduce.py
@@ -32,7 +32,7 @@
 try:
     import ray
 except ModuleNotFoundError:
-    from tensorrt_llm import ray_stub as ray
+    from tensorrt_llm.executor.ray import stub as ray
 
 from tensorrt_llm._torch.distributed.communicator import TorchDist
 from tensorrt_llm.functional import AllReduceFusionOp, AllReduceStrategy
@@ -60,7 +60,7 @@ def __init__(self, rank: int, world_size: int):
 
         assert len(ray.get_gpu_ids()) == 1
         self.gpu = int(ray.get_gpu_ids()[0])
-        from tensorrt_llm.executor.ray_gpu_worker import RayWorkerWrapper
+        from tensorrt_llm.executor.ray.gpu_worker import RayWorkerWrapper
 
         torch.cuda.set_device(RayWorkerWrapper.physical_to_local_id(self.gpu))
 
diff --git a/tests/unittest/_torch/ray_orchestrator/multi_gpu/test_refit_compile_isolation.py b/tests/unittest/_torch/ray_orchestrator/multi_gpu/test_refit_compile_isolation.py
index 40fc2674b546..0d737b90de99 100644
--- a/tests/unittest/_torch/ray_orchestrator/multi_gpu/test_refit_compile_isolation.py
+++ b/tests/unittest/_torch/ray_orchestrator/multi_gpu/test_refit_compile_isolation.py
@@ -92,8 +92,8 @@
 from utils.util import skip_pre_rubin
 
 from tensorrt_llm import LLM
-from tensorrt_llm._ray_utils import control_action_decorator
 from tensorrt_llm._torch.utils import get_device_uuid
+from tensorrt_llm.executor.ray.utils import control_action_decorator
 from tensorrt_llm.llmapi import KvCacheConfig, MoeConfig, SamplingParams, TorchCompileConfig
 from tensorrt_llm.llmapi.rlhf_utils import WorkerExtension