diff --git a/Source/Engine/Content/Storage/FlaxStorage.cpp b/Source/Engine/Content/Storage/FlaxStorage.cpp
index 17daba278..ed8b623ae 100644
--- a/Source/Engine/Content/Storage/FlaxStorage.cpp
+++ b/Source/Engine/Content/Storage/FlaxStorage.cpp
@@ -1425,6 +1425,7 @@ bool FlaxStorage::CloseFileHandles()
         return false;
     }
     PROFILE_CPU();
+    PROFILE_MEM(ContentFiles);
 
     // Note: this is usually called by the content manager when this file is not used or on exit
     // In those situations all the async tasks using this storage should be cancelled externally
diff --git a/Source/Engine/Core/Delegate.h b/Source/Engine/Core/Delegate.h
index b11d9b845..8efc2beec 100644
--- a/Source/Engine/Core/Delegate.h
+++ b/Source/Engine/Core/Delegate.h
@@ -12,6 +12,9 @@
 #include "Engine/Threading/Threading.h"
 #include "Engine/Core/Collections/HashSet.h"
 #endif
+#if COMPILE_WITH_PROFILER
+#include "Engine/Profiler/ProfilerMemory.h"
+#endif
 
 /// <summary>
 /// The function object that supports binding static, member and lambda functions.
@@ -457,6 +460,7 @@ public:
     /// <param name="f">The function to bind.</param>
     void Bind(const FunctionType& f)
     {
+        PROFILE_MEM(EngineDelegate);
 #if DELEGATE_USE_ATOMIC
         const intptr size = Platform::AtomicRead(&_size);
         FunctionType* bindings = (FunctionType*)Platform::AtomicRead(&_ptr);
diff --git a/Source/Engine/Graphics/Textures/GPUTexture.cpp b/Source/Engine/Graphics/Textures/GPUTexture.cpp
index 245fb1e01..07b85bc42 100644
--- a/Source/Engine/Graphics/Textures/GPUTexture.cpp
+++ b/Source/Engine/Graphics/Textures/GPUTexture.cpp
@@ -628,6 +628,7 @@ GPUTask* GPUTexture::UploadMipMapAsync(const BytesContainer& data, int32 mipInde
 GPUTask* GPUTexture::UploadMipMapAsync(const BytesContainer& data, int32 mipIndex, int32 rowPitch, int32 slicePitch, bool copyData)
 {
     PROFILE_CPU();
+    PROFILE_MEM(GraphicsTextures);
     ASSERT(IsAllocated());
     ASSERT(mipIndex < MipLevels() && data.IsValid());
     ASSERT(data.Length() >= slicePitch);
@@ -720,6 +721,7 @@ bool GPUTexture::DownloadData(TextureData& result)
         MISSING_CODE("support volume texture data downloading.");
     }
     PROFILE_CPU();
+    PROFILE_MEM(GraphicsTextures);
 
     // Use faster path for staging resources
     if (IsStaging()) // TODO: what about chips with unified memory? if rendering is not active then we can access GPU memory from CPU directly (eg. mobile, integrated GPUs and some consoles)
@@ -806,6 +808,7 @@ Task* GPUTexture::DownloadDataAsync(TextureData& result)
         MISSING_CODE("support volume texture data downloading.");
     }
     PROFILE_CPU();
+    PROFILE_MEM(GraphicsTextures);
 
     // Use faster path for staging resources
     if (IsStaging())
diff --git a/Source/Engine/Input/Input.cpp b/Source/Engine/Input/Input.cpp
index e0fc0297f..8438977b1 100644
--- a/Source/Engine/Input/Input.cpp
+++ b/Source/Engine/Input/Input.cpp
@@ -169,6 +169,7 @@ void Mouse::OnMouseMoved(const Float2& newPosition)
 
 void Mouse::OnMouseDown(const Float2& position, const MouseButton button, Window* target)
 {
+    PROFILE_MEM(Input);
     Event& e = _queue.AddOne();
     e.Type = EventType::MouseDown;
     e.Target = target;
@@ -187,6 +188,7 @@ bool Mouse::IsAnyButtonDown() const
 
 void Mouse::OnMouseUp(const Float2& position, const MouseButton button, Window* target)
 {
+    PROFILE_MEM(Input);
     Event& e = _queue.AddOne();
     e.Type = EventType::MouseUp;
     e.Target = target;
@@ -196,6 +198,7 @@ void Mouse::OnMouseUp(const Float2& position, const MouseButton button, Window*
 
 void Mouse::OnMouseDoubleClick(const Float2& position, const MouseButton button, Window* target)
 {
+    PROFILE_MEM(Input);
     Event& e = _queue.AddOne();
     e.Type = EventType::MouseDoubleClick;
     e.Target = target;
@@ -205,6 +208,7 @@ void Mouse::OnMouseDoubleClick(const Float2& position, const MouseButton button,
 
 void Mouse::OnMouseMove(const Float2& position, Window* target)
 {
+    PROFILE_MEM(Input);
     Event& e = _queue.AddOne();
     e.Type = EventType::MouseMove;
     e.Target = target;
@@ -213,6 +217,7 @@ void Mouse::OnMouseMove(const Float2& position, Window* target)
 
 void Mouse::OnMouseLeave(Window* target)
 {
+    PROFILE_MEM(Input);
     Event& e = _queue.AddOne();
     e.Type = EventType::MouseLeave;
     e.Target = target;
@@ -220,6 +225,7 @@ void Mouse::OnMouseLeave(Window* target)
 
 void Mouse::OnMouseWheel(const Float2& position, float delta, Window* target)
 {
+    PROFILE_MEM(Input);
     Event& e = _queue.AddOne();
     e.Type = EventType::MouseWheel;
     e.Target = target;
@@ -316,6 +322,7 @@ void Keyboard::OnCharInput(Char c, Window* target)
     if (c < 32)
         return;
 
+    PROFILE_MEM(Input);
     Event& e = _queue.AddOne();
     e.Type = EventType::Char;
     e.Target = target;
@@ -326,6 +333,7 @@ void Keyboard::OnKeyUp(KeyboardKeys key, Window* target)
 {
     if (key >= KeyboardKeys::MAX)
         return;
+    PROFILE_MEM(Input);
     Event& e = _queue.AddOne();
     e.Type = EventType::KeyUp;
     e.Target = target;
@@ -336,6 +344,7 @@ void Keyboard::OnKeyDown(KeyboardKeys key, Window* target)
 {
     if (key >= KeyboardKeys::MAX)
         return;
+    PROFILE_MEM(Input);
     Event& e = _queue.AddOne();
     e.Type = EventType::KeyDown;
     e.Target = target;
diff --git a/Source/Engine/Profiler/ProfilerCPU.cpp b/Source/Engine/Profiler/ProfilerCPU.cpp
index 9ebc39bd5..5f5141e9e 100644
--- a/Source/Engine/Profiler/ProfilerCPU.cpp
+++ b/Source/Engine/Profiler/ProfilerCPU.cpp
@@ -3,6 +3,7 @@
 #if COMPILE_WITH_PROFILER
 
 #include "ProfilerCPU.h"
+#include "ProfilerMemory.h"
 #include "Engine/Engine/Globals.h"
 #include "Engine/Threading/ThreadRegistry.h"
 
@@ -157,6 +158,7 @@ int32 ProfilerCPU::BeginEvent()
     auto thread = Thread::Current;
     if (thread == nullptr)
     {
+        PROFILE_MEM(Profiler);
         const auto id = Platform::GetCurrentThreadID();
         const auto t = ThreadRegistry::GetThread(id);
         if (t)
diff --git a/Source/Engine/Profiler/ProfilerGPU.cpp b/Source/Engine/Profiler/ProfilerGPU.cpp
index 168677c1b..9330663f4 100644
--- a/Source/Engine/Profiler/ProfilerGPU.cpp
+++ b/Source/Engine/Profiler/ProfilerGPU.cpp
@@ -3,6 +3,7 @@
 #if COMPILE_WITH_PROFILER
 
 #include "ProfilerGPU.h"
+#include "ProfilerMemory.h"
 #include "Engine/Core/Log.h"
 #include "Engine/Engine/Engine.h"
 #include "Engine/Graphics/GPUDevice.h"
@@ -45,6 +46,7 @@ void ProfilerGPU::EventBuffer::TryResolve()
     }
 
     // Collect queries results and free them
+    PROFILE_MEM(Profiler);
     for (int32 i = 0; i < _data.Count(); i++)
     {
         auto& e = _data[i];
@@ -58,6 +60,7 @@ void ProfilerGPU::EventBuffer::TryResolve()
 
 int32 ProfilerGPU::EventBuffer::Add(const Event& e)
 {
+    PROFILE_MEM(Profiler);
     const int32 index = _data.Count();
     _data.Add(e);
     return index;
@@ -88,6 +91,7 @@ GPUTimerQuery* ProfilerGPU::GetTimerQuery()
     }
     else
     {
+        PROFILE_MEM(Profiler);
         result = GPUDevice::Instance->CreateTimerQuery();
         _timerQueriesPool.Add(result);
     }
diff --git a/Source/Engine/Profiler/ProfilerMemory.cpp b/Source/Engine/Profiler/ProfilerMemory.cpp
index 972dd6646..807f397cc 100644
--- a/Source/Engine/Profiler/ProfilerMemory.cpp
+++ b/Source/Engine/Profiler/ProfilerMemory.cpp
@@ -234,6 +234,8 @@ void InitProfilerMemory(const Char* cmdLine, int32 stage)
 
     // Init hierarchy
 #define INIT_PARENT(parent, child) GroupParents[(int32)ProfilerMemory::Groups::child] = (uint8)ProfilerMemory::Groups::parent
+    INIT_PARENT(Engine, EngineThreading);
+    INIT_PARENT(Engine, EngineDelegate);
     INIT_PARENT(Malloc, MallocArena);
     INIT_PARENT(Graphics, GraphicsTextures);
     INIT_PARENT(Graphics, GraphicsRenderTargets);
diff --git a/Source/Engine/Profiler/ProfilerMemory.h b/Source/Engine/Profiler/ProfilerMemory.h
index 1b1ad5ac3..2112a407d 100644
--- a/Source/Engine/Profiler/ProfilerMemory.h
+++ b/Source/Engine/Profiler/ProfilerMemory.h
@@ -30,8 +30,6 @@ public:
         TotalUntracked,
         // Initial memory used by program upon startup (eg. executable size, static variables).
         ProgramSize,
-        // General purpose engine memory.
-        Engine,
         // Profiling tool memory overhead.
         Profiler,
 
@@ -40,6 +38,13 @@ public:
         // Total memory allocated via arena allocators (all pages).
         MallocArena,
 
+        // General purpose engine memory.
+        Engine,
+        // Memory used by the threads (and relevant systems such as Job System).
+        EngineThreading,
+        // Memory used by Delegate (engine events system to store all references).
+        EngineDelegate,
+
         // Total graphics memory usage.
         Graphics,
         // Total textures memory usage.
diff --git a/Source/Engine/Threading/ConcurrentQueue.h b/Source/Engine/Threading/ConcurrentQueue.h
index ba60dcec6..79c819cc8 100644
--- a/Source/Engine/Threading/ConcurrentQueue.h
+++ b/Source/Engine/Threading/ConcurrentQueue.h
@@ -3,6 +3,7 @@
 #pragma once
 
 #include "Engine/Core/Memory/Memory.h"
+#include "Engine/Profiler/ProfilerMemory.h"
 #define MOODYCAMEL_EXCEPTIONS_ENABLED 0
 #include <ThirdParty/concurrentqueue.h>
 
@@ -17,6 +18,7 @@ struct ConcurrentQueueSettings : public moodycamel::ConcurrentQueueDefaultTraits
     // Use default engine memory allocator
     static inline void* malloc(size_t size)
     {
+        PROFILE_MEM(EngineThreading);
         return Allocator::Allocate((uint64)size);
     }
 
diff --git a/Source/Engine/Threading/JobSystem.cpp b/Source/Engine/Threading/JobSystem.cpp
index 634f1a884..612584c40 100644
--- a/Source/Engine/Threading/JobSystem.cpp
+++ b/Source/Engine/Threading/JobSystem.cpp
@@ -12,6 +12,7 @@
 #include "Engine/Core/Collections/RingBuffer.h"
 #include "Engine/Engine/EngineService.h"
 #include "Engine/Profiler/ProfilerCPU.h"
+#include "Engine/Profiler/ProfilerMemory.h"
 #if USE_CSHARP
 #include "Engine/Scripting/ManagedCLR/MCore.h"
 #endif
@@ -118,17 +119,22 @@ void* JobSystemAllocation::Allocate(uintptr size)
         }
     }
     if (!result)
+    {
+        PROFILE_MEM(EngineThreading);
         result = Platform::Allocate(size, 16);
+    }
     return result;
 }
 
 void JobSystemAllocation::Free(void* ptr, uintptr size)
 {
+    PROFILE_MEM(EngineThreading);
     MemPool.Add({ ptr, size });
 }
 
 bool JobSystemService::Init()
 {
+    PROFILE_MEM(EngineThreading);
     ThreadsCount = Math::Min<int32>(Platform::GetCPUInfo().LogicalProcessorCount, ARRAY_COUNT(Threads));
     for (int32 i = 0; i < ThreadsCount; i++)
     {
diff --git a/Source/Engine/Threading/ThreadPool.cpp b/Source/Engine/Threading/ThreadPool.cpp
index aeb4dea98..e84aa2cdd 100644
--- a/Source/Engine/Threading/ThreadPool.cpp
+++ b/Source/Engine/Threading/ThreadPool.cpp
@@ -14,6 +14,7 @@
 #include "Engine/Platform/ConditionVariable.h"
 #include "Engine/Platform/CPUInfo.h"
 #include "Engine/Platform/Thread.h"
+#include "Engine/Profiler/ProfilerMemory.h"
 
 FLAXENGINE_API bool IsInMainThread()
 {
@@ -36,6 +37,7 @@ String ThreadPoolTask::ToString() const
 
 void ThreadPoolTask::Enqueue()
 {
+    PROFILE_MEM(EngineThreading);
     ThreadPoolImpl::Jobs.Add(this);
     ThreadPoolImpl::JobsSignal.NotifyOne();
 }
@@ -58,6 +60,8 @@ ThreadPoolService ThreadPoolServiceInstance;
 
 bool ThreadPoolService::Init()
 {
+    PROFILE_MEM(EngineThreading);
+
     // Spawn threads
     const int32 numThreads = Math::Clamp<int32>(Platform::GetCPUInfo().ProcessorCoreCount - 1, 2, PLATFORM_THREADS_LIMIT / 2);
     LOG(Info, "Spawning {0} Thread Pool workers", numThreads);
diff --git a/Source/Engine/Threading/ThreadRegistry.cpp b/Source/Engine/Threading/ThreadRegistry.cpp
index d793f7e56..522a5479b 100644
--- a/Source/Engine/Threading/ThreadRegistry.cpp
+++ b/Source/Engine/Threading/ThreadRegistry.cpp
@@ -3,10 +3,11 @@
 #include "ThreadRegistry.h"
 #include "Engine/Core/Collections/Dictionary.h"
 #include "Engine/Platform/CriticalSection.h"
+#include "Engine/Profiler/ProfilerMemory.h"
 
 namespace ThreadRegistryImpl
 {
-    Dictionary<uint64, Thread*> Registry(64);
+    Dictionary<uint64, Thread*> Registry;
     CriticalSection Locker;
 }
 
@@ -46,6 +47,7 @@ void ThreadRegistry::KillEmAll()
 
 void ThreadRegistry::Add(Thread* thread)
 {
+    PROFILE_MEM(EngineThreading);
     ASSERT(thread && thread->GetID() != 0);
     Locker.Lock();
     ASSERT(!Registry.ContainsKey(thread->GetID()) && !Registry.ContainsValue(thread));
diff --git a/Source/Engine/Tools/TextureTool/TextureTool.cpp b/Source/Engine/Tools/TextureTool/TextureTool.cpp
index 75f4da774..b892385da 100644
--- a/Source/Engine/Tools/TextureTool/TextureTool.cpp
+++ b/Source/Engine/Tools/TextureTool/TextureTool.cpp
@@ -14,6 +14,7 @@
 #include "Engine/Graphics/PixelFormatSampler.h"
 #include "Engine/Graphics/Textures/TextureData.h"
 #include "Engine/Profiler/ProfilerCPU.h"
+#include "Engine/Profiler/ProfilerMemory.h"
 
 #if USE_EDITOR
 #include "Engine/Core/Collections/Dictionary.h"
@@ -210,6 +211,7 @@ bool TextureTool::HasAlpha(const StringView& path)
 bool TextureTool::ImportTexture(const StringView& path, TextureData& textureData)
 {
     PROFILE_CPU();
+    PROFILE_MEM(GraphicsTextures);
     LOG(Info, "Importing texture from \'{0}\'", path);
     const auto startTime = DateTime::NowUTC();
 
@@ -247,6 +249,7 @@ bool TextureTool::ImportTexture(const StringView& path, TextureData& textureData
 bool TextureTool::ImportTexture(const StringView& path, TextureData& textureData, Options options, String& errorMsg)
 {
     PROFILE_CPU();
+    PROFILE_MEM(GraphicsTextures);
     LOG(Info, "Importing texture from \'{0}\'. Options: {1}", path, options.ToString());
     const auto startTime = DateTime::NowUTC();
 
@@ -296,6 +299,7 @@ bool TextureTool::ImportTexture(const StringView& path, TextureData& textureData
 bool TextureTool::ExportTexture(const StringView& path, const TextureData& textureData)
 {
     PROFILE_CPU();
+    PROFILE_MEM(GraphicsTextures);
     LOG(Info, "Exporting texture to \'{0}\'.", path);
     const auto startTime = DateTime::NowUTC();
     ImageType type;
@@ -346,6 +350,7 @@ bool TextureTool::Convert(TextureData& dst, const TextureData& src, const PixelF
         return true;
     }
     PROFILE_CPU();
+    PROFILE_MEM(GraphicsTextures);
 
 #if COMPILE_WITH_DIRECTXTEX
     return ConvertDirectXTex(dst, src, dstFormat);
@@ -375,6 +380,7 @@ bool TextureTool::Resize(TextureData& dst, const TextureData& src, int32 dstWidt
         return true;
     }
     PROFILE_CPU();
+    PROFILE_MEM(GraphicsTextures);
 #if COMPILE_WITH_DIRECTXTEX
     return ResizeDirectXTex(dst, src, dstWidth, dstHeight);
 #elif COMPILE_WITH_STB
@@ -488,6 +494,7 @@ bool TextureTool::GetImageType(const StringView& path, ImageType& type)
 bool TextureTool::Transform(TextureData& texture, const Function<void(Color&)>& transformation)
 {   
     PROFILE_CPU();
+    PROFILE_MEM(GraphicsTextures);
     auto sampler = PixelFormatSampler::Get(texture.Format);
     if (!sampler)
         return true;
diff --git a/Source/Engine/Utilities/Screenshot.cpp b/Source/Engine/Utilities/Screenshot.cpp
index 85abf19d8..9a453701d 100644
--- a/Source/Engine/Utilities/Screenshot.cpp
+++ b/Source/Engine/Utilities/Screenshot.cpp
@@ -82,6 +82,7 @@ bool CaptureScreenshot::Run()
         LOG(Warning, "Missing target render task.");
         return true;
     }
+    PROFILE_MEM(Graphics);
 
     // TODO: how about a case two or more screenshots at the same second? update counter and check files
 
@@ -147,6 +148,7 @@ void Screenshot::Capture(GPUTexture* target, const StringView& path)
         LOG(Warning, "Cannot take screenshot. Graphics device is not ready.");
         return;
     }
+    PROFILE_MEM(Graphics);
 
     // Faster path for staging textures that contents are ready to access on a CPU
     if (target->IsStaging())
@@ -211,6 +213,7 @@ void Screenshot::Capture(SceneRenderTask* target, const StringView& path)
         LOG(Warning, "Cannot take screenshot. Graphics device is not ready.");
         return;
     }
+    PROFILE_MEM(Graphics);
 
     // Create tasks
     auto saveTask = New<CaptureScreenshot>(target, path);