diff --git a/.github/workflows/pr-test-rust.yml b/.github/workflows/pr-test-rust.yml
index 8a70f415e..ddd10339e 100644
--- a/.github/workflows/pr-test-rust.yml
+++ b/.github/workflows/pr-test-rust.yml
@@ -191,10 +191,10 @@ jobs:
 
       - name: Run Python unit tests
         run: |
-          cd sgl-model-gateway
+          cd sgl-model-gateway/bindings/python
           source "$HOME/.cargo/env"
           python3 -m pip install pytest pytest-cov pytest-xdist
-          pytest -q py_test/unit --cov=sglang_router --cov-config=bindings/python/.coveragerc --cov-report=term-missing --cov-fail-under=80
+          pytest -q tests --cov=sglang_router --cov-config=.coveragerc --cov-report=term-missing --cov-fail-under=80
 
       - name: Run Python integration tests
         run: |
diff --git a/.github/workflows/release-docker-gateway.yml b/.github/workflows/release-docker-gateway.yml
index 9f0b64614..3b5e5a68f 100644
--- a/.github/workflows/release-docker-gateway.yml
+++ b/.github/workflows/release-docker-gateway.yml
@@ -29,7 +29,7 @@ jobs:
 
       - name: Build and Push
         run: |
-          version=$(cat sgl-model-gateway/bindings/python/sglang_router/version.py | cut -d'"' -f2)
+          version=$(cat sgl-model-gateway/bindings/python/src/sglang_router/version.py | cut -d'"' -f2)
           tag=v${version}
 
           docker buildx build . -f docker/gateway.Dockerfile \
diff --git a/sgl-model-gateway/bindings/python/README.md b/sgl-model-gateway/bindings/python/README.md
index 67f9e4e77..5f913e241 100644
--- a/sgl-model-gateway/bindings/python/README.md
+++ b/sgl-model-gateway/bindings/python/README.md
@@ -6,16 +6,22 @@ This directory contains the Python bindings for the SGLang Router, built using [
 
 ```
 bindings/python/
-├── src/                    # Rust source code for Python bindings
-│   └── lib.rs              # PyO3 bindings implementation
-├── sglang_router/          # Python source code
-│   ├── __init__.py
-│   ├── version.py
-│   ├── launch_server.py
-│   ├── launch_router.py
-│   ├── router.py
-│   ├── router_args.py
-│   └── mini_lb.py
+├── src/                    # Source code (src layout)
+│   ├── lib.rs              # Rust/PyO3 bindings implementation
+│   └── sglang_router/      # Python source code
+│       ├── __init__.py
+│       ├── version.py
+│       ├── launch_server.py
+│       ├── launch_router.py
+│       ├── router.py
+│       ├── router_args.py
+│       └── mini_lb.py
+├── tests/                  # Python unit tests
+│   ├── conftest.py
+│   ├── test_validation.py
+│   ├── test_arg_parser.py
+│   ├── test_router_config.py
+│   └── test_startup_sequence.py
 ├── Cargo.toml              # Rust package configuration for bindings
 ├── pyproject.toml          # Python package configuration
 ├── setup.py                # Setup configuration
@@ -51,15 +57,15 @@ pip install dist/sglang_router-*.whl
 ## Testing
 
 ```bash
-# Run Python tests
-cd sgl-model-gateway
-pytest py_test/
+# Run Python unit tests (after maturin develop)
+cd sgl-model-gateway/bindings/python
+pytest tests/
 ```
 
 ## Configuration
 
 - **pyproject.toml**: Defines package metadata, dependencies, and build configuration
-- **python-source**: Set to "." to indicate Python source is in the same directory as pyproject.toml
+- **python-source**: Set to `"src"` indicating Python source uses the src layout
 - **module-name**: `sglang_router.sglang_router_rs` - the Rust extension module name
 
 ## Notes
diff --git a/sgl-model-gateway/bindings/python/pyproject.toml b/sgl-model-gateway/bindings/python/pyproject.toml
index 918bd4175..e817986f4 100644
--- a/sgl-model-gateway/bindings/python/pyproject.toml
+++ b/sgl-model-gateway/bindings/python/pyproject.toml
@@ -39,6 +39,7 @@ dependencies = [
 [project.optional-dependencies]
 dev = [
     "requests>=2.25.0",
+    "pytest>=7.0.0",
 ]
 
 [project.scripts]
@@ -48,7 +49,16 @@ sglang-router = "sglang_router.cli:main"
 
 
 [tool.maturin]
-python-source = "."
+python-source = "src"
 module-name = "sglang_router.sglang_router_rs"
 # Exclude bindings/python/README.md to use root README only
 exclude = ["README.md"]
+
+[tool.pytest.ini_options]
+testpaths = ["tests"]
+python_files = ["test_*.py"]
+python_classes = ["Test*"]
+python_functions = ["test_*"]
+markers = [
+    "unit: mark test as a unit test (no GPU required)",
+]
diff --git a/sgl-model-gateway/bindings/python/sglang_router/__init__.py b/sgl-model-gateway/bindings/python/src/sglang_router/__init__.py
similarity index 100%
rename from sgl-model-gateway/bindings/python/sglang_router/__init__.py
rename to sgl-model-gateway/bindings/python/src/sglang_router/__init__.py
diff --git a/sgl-model-gateway/bindings/python/sglang_router/__main__.py b/sgl-model-gateway/bindings/python/src/sglang_router/__main__.py
similarity index 100%
rename from sgl-model-gateway/bindings/python/sglang_router/__main__.py
rename to sgl-model-gateway/bindings/python/src/sglang_router/__main__.py
diff --git a/sgl-model-gateway/bindings/python/sglang_router/cli.py b/sgl-model-gateway/bindings/python/src/sglang_router/cli.py
similarity index 100%
rename from sgl-model-gateway/bindings/python/sglang_router/cli.py
rename to sgl-model-gateway/bindings/python/src/sglang_router/cli.py
diff --git a/sgl-model-gateway/bindings/python/sglang_router/launch_router.py b/sgl-model-gateway/bindings/python/src/sglang_router/launch_router.py
similarity index 100%
rename from sgl-model-gateway/bindings/python/sglang_router/launch_router.py
rename to sgl-model-gateway/bindings/python/src/sglang_router/launch_router.py
diff --git a/sgl-model-gateway/bindings/python/sglang_router/launch_server.py b/sgl-model-gateway/bindings/python/src/sglang_router/launch_server.py
similarity index 100%
rename from sgl-model-gateway/bindings/python/sglang_router/launch_server.py
rename to sgl-model-gateway/bindings/python/src/sglang_router/launch_server.py
diff --git a/sgl-model-gateway/bindings/python/sglang_router/mini_lb.py b/sgl-model-gateway/bindings/python/src/sglang_router/mini_lb.py
similarity index 100%
rename from sgl-model-gateway/bindings/python/sglang_router/mini_lb.py
rename to sgl-model-gateway/bindings/python/src/sglang_router/mini_lb.py
diff --git a/sgl-model-gateway/bindings/python/sglang_router/router.py b/sgl-model-gateway/bindings/python/src/sglang_router/router.py
similarity index 100%
rename from sgl-model-gateway/bindings/python/sglang_router/router.py
rename to sgl-model-gateway/bindings/python/src/sglang_router/router.py
diff --git a/sgl-model-gateway/bindings/python/sglang_router/router_args.py b/sgl-model-gateway/bindings/python/src/sglang_router/router_args.py
similarity index 100%
rename from sgl-model-gateway/bindings/python/sglang_router/router_args.py
rename to sgl-model-gateway/bindings/python/src/sglang_router/router_args.py
diff --git a/sgl-model-gateway/bindings/python/sglang_router/version.py b/sgl-model-gateway/bindings/python/src/sglang_router/version.py
similarity index 100%
rename from sgl-model-gateway/bindings/python/sglang_router/version.py
rename to sgl-model-gateway/bindings/python/src/sglang_router/version.py
diff --git a/sgl-model-gateway/bindings/python/tests/conftest.py b/sgl-model-gateway/bindings/python/tests/conftest.py
new file mode 100644
index 000000000..31ceb7140
--- /dev/null
+++ b/sgl-model-gateway/bindings/python/tests/conftest.py
@@ -0,0 +1,14 @@
+"""
+Pytest configuration for sglang_router Python binding tests.
+
+These are unit tests that run without GPU resources or external dependencies.
+"""
+
+import pytest
+
+
+def pytest_configure(config):
+    """Configure pytest markers."""
+    config.addinivalue_line(
+        "markers", "unit: mark test as a unit test (no GPU required)"
+    )
diff --git a/sgl-model-gateway/py_test/unit/test_arg_parser.py b/sgl-model-gateway/bindings/python/tests/test_arg_parser.py
similarity index 100%
rename from sgl-model-gateway/py_test/unit/test_arg_parser.py
rename to sgl-model-gateway/bindings/python/tests/test_arg_parser.py
diff --git a/sgl-model-gateway/py_test/unit/test_router_config.py b/sgl-model-gateway/bindings/python/tests/test_router_config.py
similarity index 100%
rename from sgl-model-gateway/py_test/unit/test_router_config.py
rename to sgl-model-gateway/bindings/python/tests/test_router_config.py
diff --git a/sgl-model-gateway/py_test/unit/test_startup_sequence.py b/sgl-model-gateway/bindings/python/tests/test_startup_sequence.py
similarity index 100%
rename from sgl-model-gateway/py_test/unit/test_startup_sequence.py
rename to sgl-model-gateway/bindings/python/tests/test_startup_sequence.py
diff --git a/sgl-model-gateway/py_test/unit/test_validation.py b/sgl-model-gateway/bindings/python/tests/test_validation.py
similarity index 100%
rename from sgl-model-gateway/py_test/unit/test_validation.py
rename to sgl-model-gateway/bindings/python/tests/test_validation.py