diff --git a/mlir/include/mlir/ExecutionEngine/Float16bits.h b/mlir/include/mlir/ExecutionEngine/Float16bits.h
--- a/mlir/include/mlir/ExecutionEngine/Float16bits.h
+++ b/mlir/include/mlir/ExecutionEngine/Float16bits.h
@@ -48,5 +48,4 @@
 // Outputs a bfloat value.
 MLIR_FLOAT16_EXPORT std::ostream &operator<<(std::ostream &os, const bf16 &d);
 
-#undef MLIR_FLOAT16_EXPORT
 #endif // MLIR_EXECUTIONENGINE_FLOAT16BITS_H_
diff --git a/mlir/lib/ExecutionEngine/CMakeLists.txt b/mlir/lib/ExecutionEngine/CMakeLists.txt
--- a/mlir/lib/ExecutionEngine/CMakeLists.txt
+++ b/mlir/lib/ExecutionEngine/CMakeLists.txt
@@ -147,8 +147,15 @@
     MLIRSparseTensorEnums
     MLIRSparseTensorRuntime
     )
+  set_property(TARGET mlir_c_runner_utils PROPERTY CXX_VISIBILITY_PRESET hidden)
   set_property(TARGET mlir_c_runner_utils PROPERTY CXX_STANDARD 17)
   target_compile_definitions(mlir_c_runner_utils PRIVATE mlir_c_runner_utils_EXPORTS)
+  if(CMAKE_SYSTEM_NAME STREQUAL "Linux")
+    # Don't export symbols from link-time dependencies, these are internal
+    # implementation details.
+    # FIXME: Add a similar fix for Windows.
+    target_link_options(mlir_c_runner_utils PRIVATE "-Wl,-exclude-libs,ALL")
+  endif()
 
   add_mlir_library(mlir_runner_utils
     SHARED
diff --git a/mlir/lib/ExecutionEngine/CRunnerUtils.cpp b/mlir/lib/ExecutionEngine/CRunnerUtils.cpp
--- a/mlir/lib/ExecutionEngine/CRunnerUtils.cpp
+++ b/mlir/lib/ExecutionEngine/CRunnerUtils.cpp
@@ -14,6 +14,7 @@
 
 #include "mlir/ExecutionEngine/CRunnerUtils.h"
 #include "mlir/ExecutionEngine/Msan.h"
+#include "llvm/ADT/StringMap.h"
 
 #ifndef _WIN32
 #if defined(__FreeBSD__) || defined(__NetBSD__) || defined(__OpenBSD__)
@@ -187,4 +188,45 @@
 IMPL_STDSORT(F32, float)
 #undef IMPL_STDSORT
 
+//===----------------------------------------------------------------------===//
+// MLIR ExecutionEngine dynamic library integration.
+//===----------------------------------------------------------------------===//
+
+// Visual Studio had a bug that fails to compile nested generic lambdas
+// inside an `extern "C"` function.
+//   https://developercommunity.visualstudio.com/content/problem/475494/clexe-error-with-lambda-inside-function-templates.html
+// The bug is fixed in VS2019 16.1. Separating the declaration and definition is
+// a work around for older versions of Visual Studio.
+extern "C" MLIR_CRUNNERUTILS_EXPORT void
+__mlir_execution_engine_init(llvm::StringMap<void *> &exportSymbols);
+
+void __mlir_execution_engine_init(llvm::StringMap<void *> &exportSymbols) {
+  auto exportSymbol = [&](llvm::StringRef name, auto ptr) {
+    assert(exportSymbols.count(name) == 0 && "symbol already exists");
+    exportSymbols[name] = reinterpret_cast<void *>(ptr);
+  };
+
+  exportSymbol("memrefCopy", &memrefCopy);
+  exportSymbol("printI64", &printI64);
+  exportSymbol("printU64", &printU64);
+  exportSymbol("printF32", &printF32);
+  exportSymbol("printF64", &printF64);
+  exportSymbol("printOpen", &printOpen);
+  exportSymbol("printClose", &printClose);
+  exportSymbol("printComma", &printComma);
+  exportSymbol("printNewline", &printNewline);
+  exportSymbol("printF16", &printF16);
+  exportSymbol("printBF16", &printBF16);
+  exportSymbol("printFlops", &printFlops);
+  exportSymbol("rtclock", &rtclock);
+  exportSymbol("*rtsrand", &*rtsrand);
+  exportSymbol("rtrand", &rtrand);
+  exportSymbol("rtdrand", &rtdrand);
+  exportSymbol("_mlir_ciface_stdSortI64", &_mlir_ciface_stdSortI64);
+  exportSymbol("_mlir_ciface_stdSortF64", &_mlir_ciface_stdSortF64);
+  exportSymbol("_mlir_ciface_stdSortF32", &_mlir_ciface_stdSortF32);
+}
+
+extern "C" MLIR_CRUNNERUTILS_EXPORT void __mlir_execution_engine_destroy() {}
+
 #endif // MLIR_CRUNNERUTILS_DEFINE_FUNCTIONS
diff --git a/mlir/lib/ExecutionEngine/Float16bits.cpp b/mlir/lib/ExecutionEngine/Float16bits.cpp
--- a/mlir/lib/ExecutionEngine/Float16bits.cpp
+++ b/mlir/lib/ExecutionEngine/Float16bits.cpp
@@ -193,12 +193,12 @@
 }
 
 // Provide these to the CRunner with the local float16 knowledge.
-extern "C" void printF16(uint16_t bits) {
+extern "C" MLIR_FLOAT16_EXPORT void printF16(uint16_t bits) {
   f16 f;
   std::memcpy(&f, &bits, sizeof(f16));
   std::cout << f;
 }
-extern "C" void printBF16(uint16_t bits) {
+extern "C" MLIR_FLOAT16_EXPORT void printBF16(uint16_t bits) {
   bf16 f;
   std::memcpy(&f, &bits, sizeof(bf16));
   std::cout << f;