diff --git a/backends/qualcomm/README.md b/backends/qualcomm/README.md
index be7cd427d6..d426aff16b 100644
--- a/backends/qualcomm/README.md
+++ b/backends/qualcomm/README.md
@@ -46,7 +46,7 @@ backends/qualcomm
 |   └── python # Python interface for using QNN libraries.
 ├── builders # Codes for lowering each operators (AoT Part).
 ├── partition # QNN Partitioner (AoT Part).
-├── passes # Various passes helping lower models to QNN backend (AoT Part).
+├── _passes # Various private passes helping lower models to QNN backend (AoT Part).
 ├── python # Places to put pybind artifacts for accessing QNN APIs, structures, etc (AoT Part).
 ├── quantizer # QNN Quantizer
 ├── runtime # Here is QNN runtime responsbile for compiling a model on x64.
diff --git a/backends/qualcomm/_passes/TARGETS b/backends/qualcomm/_passes/TARGETS
new file mode 100644
index 0000000000..a824ca9f6e
--- /dev/null
+++ b/backends/qualcomm/_passes/TARGETS
@@ -0,0 +1,18 @@
+load("@fbsource//xplat/executorch/build:runtime_wrapper.bzl", "runtime")
+
+oncall("executorch")
+
+runtime.python_library(
+    name = "passes",
+    srcs = glob([
+        "*.py",
+    ]),
+    visibility = [
+        "@EXECUTORCH_CLIENTS",
+    ],
+    deps = [
+        "//executorch/backends/transforms:addmm_mm_to_linear",
+        "//executorch/exir/backend:backend_details",
+        "//executorch/exir/backend:compile_spec_schema",
+    ],
+)
diff --git a/backends/qualcomm/passes/annotate_and_quant_scalar.py b/backends/qualcomm/_passes/annotate_and_quant_scalar.py
similarity index 100%
rename from backends/qualcomm/passes/annotate_and_quant_scalar.py
rename to backends/qualcomm/_passes/annotate_and_quant_scalar.py
diff --git a/backends/qualcomm/passes/annotate_decomposed.py b/backends/qualcomm/_passes/annotate_decomposed.py
similarity index 100%
rename from backends/qualcomm/passes/annotate_decomposed.py
rename to backends/qualcomm/_passes/annotate_decomposed.py
diff --git a/backends/qualcomm/passes/annotate_quant_attrs.py b/backends/qualcomm/_passes/annotate_quant_attrs.py
similarity index 100%
rename from backends/qualcomm/passes/annotate_quant_attrs.py
rename to backends/qualcomm/_passes/annotate_quant_attrs.py
diff --git a/backends/qualcomm/passes/build_quant_io.py b/backends/qualcomm/_passes/build_quant_io.py
similarity index 100%
rename from backends/qualcomm/passes/build_quant_io.py
rename to backends/qualcomm/_passes/build_quant_io.py
diff --git a/backends/qualcomm/passes/convert_binary_op_with_scalar.py b/backends/qualcomm/_passes/convert_binary_op_with_scalar.py
similarity index 100%
rename from backends/qualcomm/passes/convert_binary_op_with_scalar.py
rename to backends/qualcomm/_passes/convert_binary_op_with_scalar.py
diff --git a/backends/qualcomm/passes/convert_bmm_to_matmul.py b/backends/qualcomm/_passes/convert_bmm_to_matmul.py
similarity index 100%
rename from backends/qualcomm/passes/convert_bmm_to_matmul.py
rename to backends/qualcomm/_passes/convert_bmm_to_matmul.py
diff --git a/backends/qualcomm/passes/convert_interpolate_with_upsample2d.py b/backends/qualcomm/_passes/convert_interpolate_with_upsample2d.py
similarity index 100%
rename from backends/qualcomm/passes/convert_interpolate_with_upsample2d.py
rename to backends/qualcomm/_passes/convert_interpolate_with_upsample2d.py
diff --git a/backends/qualcomm/passes/convert_prelu.py b/backends/qualcomm/_passes/convert_prelu.py
similarity index 100%
rename from backends/qualcomm/passes/convert_prelu.py
rename to backends/qualcomm/_passes/convert_prelu.py
diff --git a/backends/qualcomm/passes/convert_to_linear.py b/backends/qualcomm/_passes/convert_to_linear.py
similarity index 100%
rename from backends/qualcomm/passes/convert_to_linear.py
rename to backends/qualcomm/_passes/convert_to_linear.py
diff --git a/backends/qualcomm/passes/decompose_silu.py b/backends/qualcomm/_passes/decompose_silu.py
similarity index 100%
rename from backends/qualcomm/passes/decompose_silu.py
rename to backends/qualcomm/_passes/decompose_silu.py
diff --git a/backends/qualcomm/passes/fold_qdq.py b/backends/qualcomm/_passes/fold_qdq.py
similarity index 100%
rename from backends/qualcomm/passes/fold_qdq.py
rename to backends/qualcomm/_passes/fold_qdq.py
diff --git a/backends/qualcomm/passes/fuse_consecutive_transpose.py b/backends/qualcomm/_passes/fuse_consecutive_transpose.py
similarity index 100%
rename from backends/qualcomm/passes/fuse_consecutive_transpose.py
rename to backends/qualcomm/_passes/fuse_consecutive_transpose.py
diff --git a/backends/qualcomm/passes/i64_to_i32.py b/backends/qualcomm/_passes/i64_to_i32.py
similarity index 100%
rename from backends/qualcomm/passes/i64_to_i32.py
rename to backends/qualcomm/_passes/i64_to_i32.py
diff --git a/backends/qualcomm/passes/insert_io_qdq.py b/backends/qualcomm/_passes/insert_io_qdq.py
similarity index 100%
rename from backends/qualcomm/passes/insert_io_qdq.py
rename to backends/qualcomm/_passes/insert_io_qdq.py
diff --git a/backends/qualcomm/passes/insert_requantize.py b/backends/qualcomm/_passes/insert_requantize.py
similarity index 100%
rename from backends/qualcomm/passes/insert_requantize.py
rename to backends/qualcomm/_passes/insert_requantize.py
diff --git a/backends/qualcomm/passes/layout_transform.py b/backends/qualcomm/_passes/layout_transform.py
similarity index 100%
rename from backends/qualcomm/passes/layout_transform.py
rename to backends/qualcomm/_passes/layout_transform.py
diff --git a/backends/qualcomm/passes/recompose_pixel_unshuffle.py b/backends/qualcomm/_passes/recompose_pixel_unshuffle.py
similarity index 100%
rename from backends/qualcomm/passes/recompose_pixel_unshuffle.py
rename to backends/qualcomm/_passes/recompose_pixel_unshuffle.py
diff --git a/backends/qualcomm/passes/recompose_rms_norm.py b/backends/qualcomm/_passes/recompose_rms_norm.py
similarity index 100%
rename from backends/qualcomm/passes/recompose_rms_norm.py
rename to backends/qualcomm/_passes/recompose_rms_norm.py
diff --git a/backends/qualcomm/passes/reduce_dynamic_range.py b/backends/qualcomm/_passes/reduce_dynamic_range.py
similarity index 100%
rename from backends/qualcomm/passes/reduce_dynamic_range.py
rename to backends/qualcomm/_passes/reduce_dynamic_range.py
diff --git a/backends/qualcomm/passes/remove_redundancy.py b/backends/qualcomm/_passes/remove_redundancy.py
similarity index 100%
rename from backends/qualcomm/passes/remove_redundancy.py
rename to backends/qualcomm/_passes/remove_redundancy.py
diff --git a/backends/qualcomm/passes/replace_index_put_input.py b/backends/qualcomm/_passes/replace_index_put_input.py
similarity index 100%
rename from backends/qualcomm/passes/replace_index_put_input.py
rename to backends/qualcomm/_passes/replace_index_put_input.py
diff --git a/backends/qualcomm/passes/replace_inf_buffer.py b/backends/qualcomm/_passes/replace_inf_buffer.py
similarity index 100%
rename from backends/qualcomm/passes/replace_inf_buffer.py
rename to backends/qualcomm/_passes/replace_inf_buffer.py
diff --git a/backends/qualcomm/passes/utils.py b/backends/qualcomm/_passes/utils.py
similarity index 100%
rename from backends/qualcomm/passes/utils.py
rename to backends/qualcomm/_passes/utils.py
diff --git a/backends/qualcomm/qnn_preprocess.py b/backends/qualcomm/qnn_preprocess.py
index 6360b35280..417934acbd 100644
--- a/backends/qualcomm/qnn_preprocess.py
+++ b/backends/qualcomm/qnn_preprocess.py
@@ -11,15 +11,15 @@
 import executorch.backends.qualcomm.python.PyQnnManagerAdaptor as PyQnnManager
 
 import torch  # noqa: F401
-from executorch.backends.qualcomm.builders.node_visitor import get_node_visitors
-from executorch.backends.qualcomm.builders.qnn_constants import OpContextLoader
-from executorch.backends.qualcomm.passes.convert_to_linear import ConvertToLinear
-from executorch.backends.qualcomm.passes.fuse_consecutive_transpose import (
+from executorch.backends.qualcomm._passes.convert_to_linear import ConvertToLinear
+from executorch.backends.qualcomm._passes.fuse_consecutive_transpose import (
     FuseConsecutiveTranspose,
 )
-from executorch.backends.qualcomm.passes.insert_io_qdq import InsertIOQDQ
-from executorch.backends.qualcomm.passes.insert_requantize import InsertRequantize
-from executorch.backends.qualcomm.passes.layout_transform import LayoutTransform
+from executorch.backends.qualcomm._passes.insert_io_qdq import InsertIOQDQ
+from executorch.backends.qualcomm._passes.insert_requantize import InsertRequantize
+from executorch.backends.qualcomm._passes.layout_transform import LayoutTransform
+from executorch.backends.qualcomm.builders.node_visitor import get_node_visitors
+from executorch.backends.qualcomm.builders.qnn_constants import OpContextLoader
 from executorch.backends.qualcomm.utils.utils import generate_qnn_executorch_option
 from executorch.exir.backend.backend_details import (
     BackendDetails,
diff --git a/backends/qualcomm/quantizer/quantizer.py b/backends/qualcomm/quantizer/quantizer.py
index e27edf939c..b479d79ab8 100644
--- a/backends/qualcomm/quantizer/quantizer.py
+++ b/backends/qualcomm/quantizer/quantizer.py
@@ -7,12 +7,12 @@
 from typing import Callable, Dict, Optional, Sequence, Set
 
 import torch
-from executorch.backends.qualcomm.passes.decompose_silu import DecomposeSilu
-from executorch.backends.qualcomm.passes.recompose_pixel_unshuffle import (
+from executorch.backends.qualcomm._passes.decompose_silu import DecomposeSilu
+from executorch.backends.qualcomm._passes.recompose_pixel_unshuffle import (
     RecomposePixelUnshuffle,
 )
-from executorch.backends.qualcomm.passes.reduce_dynamic_range import ReduceDynamicRange
-from executorch.backends.qualcomm.passes.replace_inf_buffer import ReplaceInfBuffer
+from executorch.backends.qualcomm._passes.reduce_dynamic_range import ReduceDynamicRange
+from executorch.backends.qualcomm._passes.replace_inf_buffer import ReplaceInfBuffer
 from executorch.backends.transforms.decompose_sdpa import (
     DecomposeScaledDotProductAttention,
 )
diff --git a/backends/qualcomm/utils/constants.py b/backends/qualcomm/utils/constants.py
index 9875c9f5af..4787e89ca6 100644
--- a/backends/qualcomm/utils/constants.py
+++ b/backends/qualcomm/utils/constants.py
@@ -6,7 +6,7 @@
 
 # Qualcomm specific key
 
-# constants in backends/qualcomm/passes & backends/qualcomm/builders
+# constants in backends/qualcomm/_passes & backends/qualcomm/builders
 QCOM_AXIS = "axis"
 QCOM_AXIS_ORDER = "axis_order"
 QCOM_BITWIDTH = "bitwidth"
diff --git a/backends/qualcomm/utils/utils.py b/backends/qualcomm/utils/utils.py
index e1be24d0d6..4509843aba 100644
--- a/backends/qualcomm/utils/utils.py
+++ b/backends/qualcomm/utils/utils.py
@@ -14,37 +14,42 @@
 import executorch.exir as exir
 
 import torch
-
-from executorch.backends.qualcomm.builders.node_visitor import (
-    QNN_QUANT_TYPE_MAP,
-    QNN_TENSOR_TYPE_MAP,
-)
-from executorch.backends.qualcomm.builders.qnn_constants import OpContextLoader
-from executorch.backends.qualcomm.passes.annotate_and_quant_scalar import (
+from executorch.backends.qualcomm._passes.annotate_and_quant_scalar import (
     AnnotateAndQuantScalar,
 )
-from executorch.backends.qualcomm.passes.annotate_decomposed import AnnotateDecomposed
-from executorch.backends.qualcomm.passes.annotate_quant_attrs import AnnotateQuantAttrs
-from executorch.backends.qualcomm.passes.convert_binary_op_with_scalar import (
+from executorch.backends.qualcomm._passes.annotate_decomposed import AnnotateDecomposed
+from executorch.backends.qualcomm._passes.annotate_quant_attrs import AnnotateQuantAttrs
+from executorch.backends.qualcomm._passes.convert_binary_op_with_scalar import (
     ConvertBinaryOpsWithScalar,
 )
-from executorch.backends.qualcomm.passes.convert_bmm_to_matmul import ConvertBmmToMatmul
-from executorch.backends.qualcomm.passes.convert_interpolate_with_upsample2d import (
+from executorch.backends.qualcomm._passes.convert_bmm_to_matmul import (
+    ConvertBmmToMatmul,
+)
+from executorch.backends.qualcomm._passes.convert_interpolate_with_upsample2d import (
     ConvertInterpolateWithUpsample2D,
 )
-from executorch.backends.qualcomm.passes.convert_prelu import ConvertPReLU
-from executorch.backends.qualcomm.passes.convert_to_linear import ConvertToLinear
-from executorch.backends.qualcomm.passes.fold_qdq import FoldQDQ
-from executorch.backends.qualcomm.passes.i64_to_i32 import I64toI32
-from executorch.backends.qualcomm.passes.layout_transform import LayoutTransform
-from executorch.backends.qualcomm.passes.recompose_pixel_unshuffle import (
+from executorch.backends.qualcomm._passes.convert_prelu import ConvertPReLU
+from executorch.backends.qualcomm._passes.convert_to_linear import ConvertToLinear
+from executorch.backends.qualcomm._passes.fold_qdq import FoldQDQ
+from executorch.backends.qualcomm._passes.i64_to_i32 import I64toI32
+from executorch.backends.qualcomm._passes.layout_transform import LayoutTransform
+from executorch.backends.qualcomm._passes.recompose_pixel_unshuffle import (
     RecomposePixelUnshuffle,
 )
-from executorch.backends.qualcomm.passes.recompose_rms_norm import RecomposeRmsNorm
-from executorch.backends.qualcomm.passes.remove_redundancy import RemoveRedundancy
-from executorch.backends.qualcomm.passes.replace_index_put_input import (
+from executorch.backends.qualcomm._passes.recompose_rms_norm import RecomposeRmsNorm
+from executorch.backends.qualcomm._passes.remove_redundancy import RemoveRedundancy
+from executorch.backends.qualcomm._passes.replace_index_put_input import (
     ReplaceIndexPutInput,
 )
+
+from executorch.backends.qualcomm.builders.node_visitor import (
+    QNN_QUANT_TYPE_MAP,
+    QNN_TENSOR_TYPE_MAP,
+)
+from executorch.backends.qualcomm.builders.qnn_constants import OpContextLoader
+from executorch.backends.qualcomm.passes.expand_broadcast_tensor_shape import (
+    ExpandBroadcastTensorShape,
+)
 from executorch.backends.qualcomm.serialization.qnn_compile_spec_schema import (
     _soc_info_table,
     QcomChipset,
diff --git a/examples/qualcomm/oss_scripts/llama2/llama.py b/examples/qualcomm/oss_scripts/llama2/llama.py
index 9712b1c08f..cd7ec24d3a 100644
--- a/examples/qualcomm/oss_scripts/llama2/llama.py
+++ b/examples/qualcomm/oss_scripts/llama2/llama.py
@@ -12,9 +12,9 @@
 from multiprocessing.connection import Client
 
 import torch
+from executorch.backends.qualcomm._passes.build_quant_io import BuildQuantIo
 
 from executorch.backends.qualcomm.partition.qnn_partitioner import QnnPartitioner
-from executorch.backends.qualcomm.passes.build_quant_io import BuildQuantIo
 
 from executorch.backends.qualcomm.quantizer.quantizer import QuantDtype
 from executorch.backends.qualcomm.serialization.qnn_compile_spec_schema import (