diff --git a/c10/core/ConstantSymNodeImpl.h b/c10/core/ConstantSymNodeImpl.h
index 0e8c4a03de9e..931265cd6fcf 100644
--- a/c10/core/ConstantSymNodeImpl.h
+++ b/c10/core/ConstantSymNodeImpl.h
@@ -1,5 +1,5 @@
 #include <c10/core/SymNodeImpl.h>
-#include <c10/util/variant.h>
+#include <variant>
 
 namespace c10 {
 
@@ -39,11 +39,11 @@ class C10_API ConstantSymNodeImpl : public SymNodeImpl {
   }
   int64_t int_() override {
     TORCH_CHECK(is_int(), "not an int");
-    return c10::get<int64_t>(value_);
+    return std::get<int64_t>(value_);
   }
   bool bool_() override {
     TORCH_CHECK(is_bool(), "not a bool");
-    return c10::get<bool>(value_);
+    return std::get<bool>(value_);
   }
   bool has_hint() override {
     return true;
@@ -56,21 +56,21 @@ class C10_API ConstantSymNodeImpl : public SymNodeImpl {
   c10::SymNode gt(const c10::SymNode& other) override;
   std::string str() override {
     if constexpr (is_int_()) {
-      return std::to_string(c10::get<int64_t>(value_));
+      return std::to_string(std::get<int64_t>(value_));
     } else {
-      return c10::get<bool>(value_) ? "true" : "false";
+      return std::get<bool>(value_) ? "true" : "false";
     }
   }
   c10::optional<int64_t> constant_int() override {
     if constexpr (is_int_()) {
-      return c10::get<int64_t>(value_);
+      return std::get<int64_t>(value_);
     } else {
       return c10::nullopt;
     }
   }
   c10::optional<bool> constant_bool() override {
     if constexpr (is_bool_()) {
-      return c10::get<bool>(value_);
+      return std::get<bool>(value_);
     } else {
       return c10::nullopt;
     }
@@ -83,7 +83,7 @@ class C10_API ConstantSymNodeImpl : public SymNodeImpl {
   }
 
  private:
-  c10::variant<int64_t, bool> value_;
+  std::variant<int64_t, bool> value_;
 
   static constexpr bool is_int_() {
     return std::is_same<T, int64_t>::value;
diff --git a/torch/csrc/autograd/profiler_kineto.cpp b/torch/csrc/autograd/profiler_kineto.cpp
index d7d340f8b49b..3021458c8c05 100644
--- a/torch/csrc/autograd/profiler_kineto.cpp
+++ b/torch/csrc/autograd/profiler_kineto.cpp
@@ -100,7 +100,7 @@ auto parseArgData(
   std::vector<c10::IValue> concrete_inputs_list;
 
   for (const auto& i : c10::irange(input_shapes.size())) {
-    c10::visit(
+    std::visit(
         c10::overloaded(
             [&](const TensorMetadata& t) {
               shapes[i] = t.sizes_;
@@ -127,12 +127,12 @@ auto parseArgData(
     concrete_inputs_list.resize(input_shapes.size());
 
     for (const auto& i : c10::irange(input_shapes.size())) {
-      c10::visit(
+      std::visit(
           c10::overloaded(
               [&](const c10::IValue& val) { concrete_inputs_list[i] = val; },
               [&](const auto&) {}),
           input_shapes[i]);
-      c10::visit(
+      std::visit(
           c10::overloaded(
               [&](const c10::IValue& val) {
                 concrete_inputs_list[i] = val;
diff --git a/torch/csrc/profiler/collection.h b/torch/csrc/profiler/collection.h
index 20ef0c85836e..f998c88b15c9 100644
--- a/torch/csrc/profiler/collection.h
+++ b/torch/csrc/profiler/collection.h
@@ -5,6 +5,7 @@
 #include <mutex>
 #include <type_traits>
 #include <utility>
+#include <variant>
 
 #include <ATen/Context.h>
 #include <c10/core/Device.h>
@@ -85,7 +86,7 @@ struct TORCH_API TensorMetadata : public RawTensorMetadataBase {
   c10::optional<AllocationID> allocation_id_;
 };
 
-using op_input_t = c10::variant<
+using op_input_t = std::variant<
     TensorMetadata,
     std::vector<TensorMetadata>,
     c10::IValue,
diff --git a/torch/csrc/profiler/data_flow.cpp b/torch/csrc/profiler/data_flow.cpp
index dcb3eaffd439..d2c5b6a768e6 100644
--- a/torch/csrc/profiler/data_flow.cpp
+++ b/torch/csrc/profiler/data_flow.cpp
@@ -77,7 +77,7 @@ void calculateUniqueTensorIDs(
       result->visit(c10::overloaded(
           [&](ExtraFields<EventType::TorchOp>& torch_op) {
             for (auto& i : torch_op.inputs_) {
-              c10::visit(raw_tensors, i);
+              std::visit(raw_tensors, i);
             }
           },
           [&](ExtraFields<EventType::PyCall>& py_call) {
diff --git a/torch/csrc/profiler/python/init.cpp b/torch/csrc/profiler/python/init.cpp
index 14eb91c64f88..33b3e0f01f64 100644
--- a/torch/csrc/profiler/python/init.cpp
+++ b/torch/csrc/profiler/python/init.cpp
@@ -37,7 +37,7 @@ static void THPCapturedTraceback_dealloc(PyObject* self_) {
 
 PyTypeObject THPCapturedTracebackType = {
     PyVarObject_HEAD_INIT(
-        NULL,
+        nullptr,
         0) "torch._C._profiler.CapturedTraceback", /* tp_name */
     sizeof(THPCapturedTraceback), /* tp_basicsize */
     0, /* tp_itemsize */
@@ -162,7 +162,12 @@ int RecordFunctionFast_init(
   constexpr const char* kwlist[] = {"name", nullptr};
   PyObject* name = nullptr;
   if (!PyArg_ParseTupleAndKeywords(
-          args, kwargs, "O", const_cast<char**>(kwlist), &name)) {
+          args,
+          kwargs,
+          "O",
+          // NOLINTNEXTLINE(cppcoreguidelines-pro-type-const-cast)
+          const_cast<char**>(kwlist),
+          &name)) {
     return -1;
   }
   if (name) {
@@ -383,7 +388,7 @@ void initPythonBindings(PyObject* module) {
           [](const torch_op_t& op) {
             py::list out;
             for (const auto& input : op.inputs_) {
-              c10::visit(
+              std::visit(
                   c10::overloaded(
                       [&](const c10::IValue& v) {
                         out.append(torch::jit::toPyObject(v));
@@ -536,11 +541,11 @@ void initPythonBindings(PyObject* module) {
   static PyMethodDef RecordFunctionFast_methods[] = {
       {"__enter__", RecordFunctionFast_enter, METH_NOARGS, nullptr},
       {"__exit__", RecordFunctionFast_exit, METH_VARARGS, nullptr},
-      {NULL},
+      {nullptr},
   };
 
   static PyTypeObject RecordFunctionFast_Type = {
-      PyVarObject_HEAD_INIT(NULL, 0)};
+      PyVarObject_HEAD_INIT(nullptr, 0)};
 
   RecordFunctionFast_Type.tp_name = "torch._C._profiler.RecordFunctionFast",
   RecordFunctionFast_Type.tp_basicsize = sizeof(RecordFunctionFast);