diff --git a/loopy/target/numba.py b/loopy/target/numba.py
index 6946063ee04f52a4890344b4cbff9446bacb6923..81614371079e1073049530d9fce217c790157d92 100644
--- a/loopy/target/numba.py
+++ b/loopy/target/numba.py
@@ -93,6 +93,7 @@ class NumbaTarget(TargetBase):
         from warnings import warn
         warn("The Numba targets are not yet feature-complete",
                 LoopyWarning, stacklevel=2)
+        self._jit_func_cache = {}
 
     def split_kernel_at_global_barriers(self):
         return False
@@ -103,6 +104,18 @@ class NumbaTarget(TargetBase):
     def get_device_ast_builder(self):
         return NumbaJITASTBuilder(self)
 
+    def get_kernel_executor_cache_key(self, *args, **kwargs):
+        return None
+
+    def get_kernel_executor(self, knl, *args, **kwargs):
+        if knl not in self._jit_func_cache:
+            from loopy import generate_code
+            code, _ = generate_code(knl)
+            ns = {}
+            exec(code, ns)
+            self._jit_func_cache[knl] = ns[knl.name]
+        return self._jit_func_cache[knl]
+
     # {{{ types
 
     @memoize_method
diff --git a/test/test_target.py b/test/test_target.py
index b656383e7bbe008892f45159faadd2d195d67a3b..046e6df5724afb4ee0d6d2ebb15bef575910bde4 100644
--- a/test/test_target.py
+++ b/test/test_target.py
@@ -224,6 +224,20 @@ def test_numba_cuda_target():
     print(lp.generate_code_v2(knl).all_code())
 
 
+def test_numba_call():
+    try:
+        import numba  # noqa
+    except ImportError:
+        pytest.skip("Numba not available.")
+    target = lp.NumbaTarget()
+    knl = lp.make_kernel("{ [i]: 0<=i<n }", "out[i] = 2*a[i]", target=target)
+    typed = lp.add_dtypes(knl, {'a': np.float32})
+    a, out = np.zeros((2, 10), np.float32)
+    a[:] = np.r_[:a.size]
+    typed(a, 10, out)
+    np.testing.assert_allclose(out, a * 2)
+
+
 if __name__ == "__main__":
     if len(sys.argv) > 1:
         exec(sys.argv[1])