diff --git a/examples/scripts/code_runner.py b/examples/scripts/code_runner.py
index 3870b0fb..7646455b 100644
--- a/examples/scripts/code_runner.py
+++ b/examples/scripts/code_runner.py
@@ -570,7 +570,9 @@ def run(self) -> None:
         runtime_dir = os.path.join(self.project_root, "src", "runtime", self.runtime_name, "runtime")
         runtime_include_dirs.append(runtime_dir)
 
-        for kernel in self.kernels:
+        from concurrent.futures import ThreadPoolExecutor
+
+        def _compile_one_kernel(kernel):
             logger.info(f"Compiling kernel: {kernel['source']} (func_id={kernel['func_id']})")
             incore_o = kernel_compiler.compile_incore(
                 kernel["source"],
@@ -584,8 +586,10 @@ def run(self) -> None:
                 kernel_bin = incore_o  # Complete .so for dlopen
             else:
                 kernel_bin = extract_text_section(incore_o)  # .text only for mmap
+            return (kernel["func_id"], kernel_bin)
 
-            kernel_binaries.append((kernel["func_id"], kernel_bin))
+        with ThreadPoolExecutor(max_workers=len(self.kernels)) as executor:
+            kernel_binaries = list(executor.map(_compile_one_kernel, self.kernels))
 
         logger.info(f"Compiled {len(kernel_binaries)} kernel(s)")
 
diff --git a/python/kernel_compiler.py b/python/kernel_compiler.py
index e9f31a99..f4c78b26 100644
--- a/python/kernel_compiler.py
+++ b/python/kernel_compiler.py
@@ -148,19 +148,7 @@ def _run_subprocess(
         label: str,
         error_hint: str = "Compiler not found"
     ) -> subprocess.CompletedProcess:
-        """Run a subprocess command with standardized logging and error handling.
-
-        Args:
-            cmd: Command and arguments
-            label: Label for log messages (e.g., "Incore", "Orchestration")
-            error_hint: Message for FileNotFoundError
-
-        Returns:
-            CompletedProcess on success
-
-        Raises:
-            RuntimeError: If command fails or executable not found
-        """
+        """Run a subprocess command with standardized logging and error handling."""
         try:
             result = subprocess.run(cmd, capture_output=True, text=True)
 
diff --git a/python/runtime_builder.py b/python/runtime_builder.py
index 715bbda5..495ca3ee 100644
--- a/python/runtime_builder.py
+++ b/python/runtime_builder.py
@@ -1,6 +1,7 @@
 import importlib.util
 import logging
 from pathlib import Path
+from concurrent.futures import ThreadPoolExecutor
 from runtime_compiler import RuntimeCompiler
 from kernel_compiler import KernelCompiler
 
@@ -83,26 +84,30 @@ def build(self, name: str) -> tuple:
 
         compiler = self._runtime_compiler
 
-        # Compile AICore kernel
-        logger.info("[1/3] Compiling AICore kernel...")
+        # Prepare configs for all three targets
         aicore_cfg = build_config["aicore"]
         aicore_include_dirs = [str((config_dir / p).resolve()) for p in aicore_cfg["include_dirs"]]
         aicore_source_dirs = [str((config_dir / p).resolve()) for p in aicore_cfg["source_dirs"]]
-        aicore_binary = compiler.compile("aicore", aicore_include_dirs, aicore_source_dirs)
 
-        # Compile AICPU kernel
-        logger.info("[2/3] Compiling AICPU kernel...")
         aicpu_cfg = build_config["aicpu"]
         aicpu_include_dirs = [str((config_dir / p).resolve()) for p in aicpu_cfg["include_dirs"]]
         aicpu_source_dirs = [str((config_dir / p).resolve()) for p in aicpu_cfg["source_dirs"]]
-        aicpu_binary = compiler.compile("aicpu", aicpu_include_dirs, aicpu_source_dirs)
 
-        # Compile Host runtime
-        logger.info("[3/3] Compiling Host runtime...")
         host_cfg = build_config["host"]
         host_include_dirs = [str((config_dir / p).resolve()) for p in host_cfg["include_dirs"]]
         host_source_dirs = [str((config_dir / p).resolve()) for p in host_cfg["source_dirs"]]
-        host_binary = compiler.compile("host", host_include_dirs, host_source_dirs)
+
+        # Compile all three targets in parallel
+        logger.info("Compiling AICore, AICPU, Host in parallel...")
+
+        with ThreadPoolExecutor(max_workers=3) as executor:
+            fut_aicore = executor.submit(compiler.compile, "aicore", aicore_include_dirs, aicore_source_dirs)
+            fut_aicpu = executor.submit(compiler.compile, "aicpu", aicpu_include_dirs, aicpu_source_dirs)
+            fut_host = executor.submit(compiler.compile, "host", host_include_dirs, host_source_dirs)
+
+            aicore_binary = fut_aicore.result()
+            aicpu_binary = fut_aicpu.result()
+            host_binary = fut_host.result()
 
         logger.info("Build complete!")
         return (host_binary, aicpu_binary, aicore_binary)
diff --git a/python/runtime_compiler.py b/python/runtime_compiler.py
index 97c3bea8..dbfecd15 100644
--- a/python/runtime_compiler.py
+++ b/python/runtime_compiler.py
@@ -6,6 +6,7 @@
 from typing import List
 from toolchain import Toolchain, CCECToolchain, Aarch64GxxToolchain, GxxToolchain
 import env_manager
+import multiprocessing
 
 logger = logging.getLogger(__name__)
 
@@ -250,7 +251,7 @@ def _run_compilation(
             cmake_cmd = ["cmake", cmake_source_dir] + cmake_args
             self._run_build_step(cmake_cmd, build_dir, platform, "CMake configuration")
 
-            make_cmd = ["make", "VERBOSE=1"]
+            make_cmd = ["make", f"-j{min(multiprocessing.cpu_count(), 32)}", "VERBOSE=1"]
             self._run_build_step(make_cmd, build_dir, platform, "Make build")
 
             # Read the compiled binary