feat(memory): add LLM-driven memory extraction from conversations

- memory/extractor: async LLM-based memory extraction from chat history
- Parses JSON lines from LLM response as memory entries
- Wire into orchestrator pipeline via PostProcess hook
- Non-blocking: runs in goroutine with 30s timeout
- Automatic dedup via existing keyword overlap logic

backend/services/ai-core/orchestrator/orchestrator.go

@@ -15,14 +15,20 @@ import (
 	"git.yeij.top/AskaEth/Cyrene_v2/services/ai-core/subsession"
 )
 
+// MemoryExtractor is the interface for async memory extraction.
+type MemoryExtractor interface {
+	ExtractFromConversation(ctx context.Context, userID string, conversation []model.Message)
+}
+
 // Pipeline orchestrates the full dialogue flow.
 type Pipeline struct {
-	llmSelector   *llm.ModelSelector
-	cache         *llm.SemanticCache
-	ctxStore      *ctxpkg.Store
-	ctxBuilder    *ctxpkg.Builder
-	subMgr        *subsession.Manager
-	personaInjector *persona.Injector
+	llmSelector      *llm.ModelSelector
+	cache            *llm.SemanticCache
+	ctxStore         *ctxpkg.Store
+	ctxBuilder       *ctxpkg.Builder
+	subMgr           *subsession.Manager
+	personaInjector  *persona.Injector
+	memoryExtractor  MemoryExtractor
 }
 
 // NewPipeline creates a new dialogue pipeline.
@@ -43,6 +49,11 @@ func NewPipeline(
 	}
 }
 
+// SetMemoryExtractor injects the memory extraction capability.
+func (p *Pipeline) SetMemoryExtractor(e MemoryExtractor) {
+	p.memoryExtractor = e
+}
+
 // ProcessInput processes a user message through the full pipeline.
 // Returns a channel of StreamToken for the LLM response.
 func (p *Pipeline) ProcessInput(ctx context.Context, sessionID, userMessage string) (<-chan model.StreamToken, *model.Intent, error) {
@@ -87,6 +98,14 @@ func (p *Pipeline) ProcessInput(ctx context.Context, sessionID, userMessage stri
 	return tokenCh, intent, nil
 }
 
+// PostProcess triggers async memory extraction after a conversation round completes.
+func (p *Pipeline) PostProcess(sessionID, userID string) {
+	if p.memoryExtractor == nil { return }
+	history := p.ctxStore.Get(sessionID)
+	if len(history) == 0 { return }
+	p.memoryExtractor.ExtractFromConversation(nil, userID, history)
+}
+
 // --- Intent Analysis ---
 
 // controlWords are common IoT control verbs.