diff --git a/app/CMakeLists.txt b/app/CMakeLists.txt
index a5c4dce..3ef8013 100644
--- a/app/CMakeLists.txt
+++ b/app/CMakeLists.txt
@@ -1 +1,2 @@
 add_subdirectory(example)
+add_subdirectory(layer_example)
diff --git a/app/layer_example/CMakeLists.txt b/app/layer_example/CMakeLists.txt
new file mode 100644
index 0000000..e88aaa6
--- /dev/null
+++ b/app/layer_example/CMakeLists.txt
@@ -0,0 +1,11 @@
+set(ARM_DIR "${CMAKE_SOURCE_DIR}/3rdparty/ComputeLibrary")
+
+add_executable(ConvolutionLayer ConvolutionLayer.cpp)
+add_executable(ElementwiseLayer ElementwiseLayer.cpp)
+
+include_directories(${ARM_DIR})
+include_directories(${ARM_DIR/include})
+link_directories(${ARM_DIR}/build)
+
+target_link_libraries(ConvolutionLayer arm_compute)
+target_link_libraries(ElementwiseLayer arm_compute)
diff --git a/app/layer_example/ConvolutionLayer.cpp b/app/layer_example/ConvolutionLayer.cpp
new file mode 100644
index 0000000..c1d90b9
--- /dev/null
+++ b/app/layer_example/ConvolutionLayer.cpp
@@ -0,0 +1,40 @@
+#include "arm_compute/runtime/NEON/NEFunctions.h"
+#include "utils/Utils.h"
+
+#include <iostream>
+using namespace arm_compute;
+using namespace utils;
+
+int main(){
+    Tensor input;
+    Tensor weight;
+    Tensor bias;
+    Tensor output;
+
+    const unsigned int N = 1;
+    const unsigned int Hin = 3;
+    const unsigned int Win = 3;
+    const unsigned int Cin = 1;
+
+    const unsigned int Hf = 3;
+    const unsigned int Wf = 3;
+
+    const unsigned int Hout = Hin - Hf + 1;
+    const unsigned int Wout = Win - Wf + 1;
+    const unsigned int Cout = 1;
+
+    input.allocator()->init(TensorInfo(TensorShape(Hin, Win, Cin), 1, DataType::F32));
+    weight.allocator()->init(TensorInfo(TensorShape(Hf, Wf, Cin, Cout), 1, DataType::F32));
+    output.allocator()->init(TensorInfo(TensorShape(Hout, Wout, Cout), 1, DataType::F32));
+
+    input.allocator()->allocate();
+    weight.allocator()->allocate();
+    output.allocator()->allocate();
+
+    NEConvolutionLayer conv;
+    conv.configure(&input, &weight, nullptr, &output, PadStrideInfo(1, 1, 0, 0));
+
+    conv.run();
+
+    output.print(std::cout);
+}
diff --git a/app/layer_example/ElementwiseLayer.cpp b/app/layer_example/ElementwiseLayer.cpp
new file mode 100644
index 0000000..05a60f6
--- /dev/null
+++ b/app/layer_example/ElementwiseLayer.cpp
@@ -0,0 +1,31 @@
+#include "arm_compute/runtime/NEON/functions/NEElementwiseOperations.h"
+#include "utils/Utils.h"
+
+#include <iostream>
+using namespace arm_compute;
+using namespace utils;
+
+int main() {
+    const int input_width = 5;
+    const int input_height = 5;
+
+    Tensor input1, input2, output;
+
+    input1.allocator()->init(TensorInfo(TensorShape(input_width, input_height, 1), 1, DataType::F32));
+    input2.allocator()->init(TensorInfo(TensorShape(input_width, input_height, 1), 1, DataType::F32));
+    output.allocator()->init(TensorInfo(TensorShape(input_width, input_height, 1), 1, DataType::F32));
+
+    input1.allocator()->allocate();
+    input2.allocator()->allocate();
+    output.allocator()->allocate();
+
+    fill_random_tensor(input1, 0.f, 1.f);
+    fill_random_tensor(input2, 0.f, 1.f);
+
+    NEElementwiseSquaredDiff elementwise;
+    elementwise.configure(&input1, &input2, &output);
+
+    elementwise.run();
+
+    output.print(std::cout);
+}