diff --git a/app/ai/skill_gap_analyzer/analyzer.py b/app/ai/skill_gap_analyzer/analyzer.py
index 5a855f8..44d7c0b 100644
--- a/app/ai/skill_gap_analyzer/analyzer.py
+++ b/app/ai/skill_gap_analyzer/analyzer.py
@@ -5,6 +5,7 @@
 """
 
 import asyncio
+import time
 
 from langchain_core.output_parsers import StrOutputParser
 from langchain_core.prompts import ChatPromptTemplate
@@ -44,20 +45,23 @@ async def analyze_skill_gap(skill_tags: list[str], resume_json: str) -> list[str
 
 _summary_optimize_chain = (
     ChatPromptTemplate.from_messages([("system", SUMMARY_OPTIMIZE_PROMPT), ("human", "请开始优化。")])
-    | LLM.JIAYU_CLAUDE_SONNET_4_5.create(temperature=0.3)
+    | LLM.DOUBAO_PRO_32K.create(temperature=0.3)
     | StrOutputParser()
 )
 
 
 async def optimize_summary(job_title: str, add_skills: list[str], original_summary: str) -> str:
     """优化个人概述，融入技能关键词"""
+    t0 = time.monotonic()
     try:
-        return await _summary_optimize_chain.ainvoke({
+        result = await _summary_optimize_chain.ainvoke({
             "job_title": job_title, "add_skills": "、".join(add_skills) if add_skills else "无",
             "original_summary": original_summary or "暂无",
         })
+        log.info(f"AI优化summary完成 ({round(time.monotonic() - t0, 2)}s)")
+        return result
     except Exception as e:
-        log.warning(f"AI优化summary失败: {e}")
+        log.warning(f"AI优化summary失败: {e} ({round(time.monotonic() - t0, 2)}s)")
         return original_summary
 
 
@@ -65,21 +69,24 @@ async def optimize_summary(job_title: str, add_skills: list[str], original_summa
 
 _experience_optimize_chain = (
     ChatPromptTemplate.from_messages([("system", EXPERIENCE_OPTIMIZE_PROMPT), ("human", "请开始优化。")])
-    | LLM.JIAYU_CLAUDE_SONNET_4_5.create(temperature=0.3)
+    | LLM.DOUBAO_PRO_32K.create(temperature=0.3)
     | StrOutputParser()
 )
 
 
 async def optimize_module(job_title: str, job_description: str, module_data: str) -> list | dict | None:
-    """优化子表模块经历描述，返回修改后的完整模块数据"""
+    """优化单条经历描述，返回修改后的记录数据"""
+    t0 = time.monotonic()
     try:
         raw = await _experience_optimize_chain.ainvoke({
             "job_title": job_title, "job_description": job_description or "",
             "original_module_data": module_data,
         })
-        return parse_llm_json(raw)
+        result = parse_llm_json(raw)
+        log.info(f"AI优化经历模块完成 ({round(time.monotonic() - t0, 2)}s)")
+        return result
     except Exception as e:
-        log.warning(f"AI优化经历模块失败: {e}")
+        log.warning(f"AI优化经历模块失败: {e} ({round(time.monotonic() - t0, 2)}s)")
         return None
 
 
diff --git a/app/ai/skill_gap_analyzer/prompts.py b/app/ai/skill_gap_analyzer/prompts.py
index d793ea5..a20a4f4 100644
--- a/app/ai/skill_gap_analyzer/prompts.py
+++ b/app/ai/skill_gap_analyzer/prompts.py
@@ -36,13 +36,13 @@ SUMMARY_OPTIMIZE_PROMPT = """你是一个简历优化助手。根据目标岗位
 4. 避免过度优化，改动越少越好
 5. 直接输出优化后的文本，不要其他内容"""
 
-EXPERIENCE_OPTIMIZE_PROMPT = """你是一个简历优化助手。根据目标岗位信息，微调用户的经历描述。
+EXPERIENCE_OPTIMIZE_PROMPT = """你是一个简历优化助手。根据目标岗位信息，微调用户的一条经历描述。
 
 【目标岗位】
 {job_title}
 {job_description}
 
-【原始经历数据】
+【原始经历数据（单条记录）】
 {original_module_data}
 
 规则：
@@ -50,7 +50,7 @@ EXPERIENCE_OPTIMIZE_PROMPT = """你是一个简历优化助手。根据目标岗
 2. 让描述更贴合目标岗位方向，但不要编造内容
 3. 避免过度优化，改动越少越好
 4. description 字段是 [{{"id": "xxx", "text": "xxx"}}] 格式：修改时保留原 id 只改 text，新增段落生成随机8位字符串作为 id，删除段落直接移除
-5. 返回修改后的完整模块数据（JSON 格式，与输入格式一致）"""
+5. 返回修改后的单条记录（JSON 对象格式，与输入格式一致，不要包裹在数组中）"""
 
 AGENT_PLAN_PROMPT = """你是一个简历编辑助手。你的唯一职责是根据用户指令修改简历内容。
 
diff --git a/app/services/skill_gap_service.py b/app/services/skill_gap_service.py
index 4c071e6..1240af0 100644
--- a/app/services/skill_gap_service.py
+++ b/app/services/skill_gap_service.py
@@ -8,6 +8,7 @@
 
 import asyncio
 import json
+import time
 
 from sqlalchemy import select
 from sqlalchemy.ext.asyncio import AsyncSession
@@ -97,7 +98,7 @@ class SkillGapService:
 
     async def generate_customize_resume(self, user_id: int, job_id: int, resume_id: int,
                                         optimize_modules: list[str], add_skills: list[str]) -> None:
-        """生成定制简历：查数据 → 并发AI优化 → 存Redis"""
+        """生成定制简历：查数据 → 按单条记录并发AI优化 → 存数据库"""
         if not optimize_modules:
             raise ValueError("请至少选择一个优化模块")
         # 1. 查简历 + 岗位
@@ -105,55 +106,71 @@ class SkillGapService:
         job = await self._get_job(job_id)
         # 2. 组装基础定制简历
         cr = customize_resume_store.build_from_detail(detail)
-        # 3. 并发 AI 优化
-        tasks = []
+        # 3. 构建并发任务（按单条记录粒度）
         job_desc = f"{job.description or ''}\n{job.requirement or ''}"
+        tasks: list[tuple[str, int, object]] = []
         if "summary" in optimize_modules:
-            tasks.append(("summary", optimize_summary(job.title or "", add_skills, detail.resume.summary or "")))
+            tasks.append(("summary", 0, optimize_summary(job.title or "", add_skills, detail.resume.summary or "")))
         if "experience" in optimize_modules:
-            for module_name, rows_json in self._experience_tasks(cr, job.title or "", job_desc):
-                tasks.append((module_name, optimize_module(job.title or "", job_desc, rows_json)))
-        # 执行并发
+            for mod_name, idx, record_json in self._experience_tasks(cr):
+                tasks.append((mod_name, idx, optimize_module(job.title or "", job_desc, record_json)))
+        log.info(f"定制简历优化开始: {len(tasks)}个并发任务 [modules={optimize_modules}, job={job_id}, resume={resume_id}]")
+        # 4. 并发执行
         if tasks:
-            keys = [t[0] for t in tasks]
-            results = await asyncio.gather(*[t[1] for t in tasks], return_exceptions=True)
-            for key, result in zip(keys, results):
-                if isinstance(result, Exception):
-                    log.warning(f"定制简历优化[{key}]失败: {result}")
+            t0 = time.monotonic()
+            results = await asyncio.gather(*[t[2] for t in tasks], return_exceptions=True)
+            log.info(f"定制简历优化全部完成, 总耗时={round(time.monotonic() - t0, 2)}s")
+            for (mod_name, idx, _), result in zip(tasks, results):
+                if isinstance(result, Exception) or result is None:
                     continue
-                self._apply_optimize_result(cr, key, result)
-        # 4. skills 追加（纯内存操作）
+                self._apply_optimize_result(cr, mod_name, idx, result)
+        # 5. skills 追加
         if "skills" in optimize_modules and add_skills:
             existing = set(cr.resume.skills)
-            cr.resume.skills.extend([s for s in add_skills if s not in existing])
-        # 5. 存数据库
+            new_skills = [s for s in add_skills if s not in existing]
+            cr.resume.skills.extend(new_skills)
+            if new_skills:
+                log.info(f"定制简历追加技能: {new_skills}")
+        # 6. 存数据库
         await customize_resume_store.save(user_id, job_id, cr)
+        log.info(f"定制简历已保存 [user={user_id}, job={job_id}]")
 
     @staticmethod
-    def _experience_tasks(cr: CustomizeResume, job_title: str, job_desc: str) -> list[tuple[str, str]]:
-        """构建各子表的 AI 优化任务列表"""
-        result = []
+    def _experience_tasks(cr: CustomizeResume) -> list[tuple[str, int, str]]:
+        """构建各子表的 AI 优化任务列表，按单条记录拆分"""
+        result: list[tuple[str, int, str]] = []
         for name, items in [("education", cr.education), ("work", cr.work), ("internship", cr.internship),
                             ("project", cr.project), ("competition", cr.competition)]:
-            if items:
-                result.append((name, json.dumps([item.model_dump(by_alias=True) for item in items], ensure_ascii=False)))
+            for idx, item in enumerate(items or []):
+                result.append((name, idx, json.dumps(item.model_dump(by_alias=True), ensure_ascii=False)))
         return result
 
     @staticmethod
-    def _apply_optimize_result(cr: CustomizeResume, key: str, result) -> None:
-        """将 AI 优化结果应用到定制简历"""
+    def _apply_optimize_result(cr: CustomizeResume, key: str, idx: int, result) -> None:
+        """将 AI 优化结果应用到定制简历（单条记录粒度）"""
         if key == "summary" and isinstance(result, str):
             cr.resume.summary = result
-        elif key == "education" and isinstance(result, list):
-            cr.education = [Education.model_validate(item) for item in result]
-        elif key == "work" and isinstance(result, list):
-            cr.work = [Work.model_validate(item) for item in result]
-        elif key == "internship" and isinstance(result, list):
-            cr.internship = [Internship.model_validate(item) for item in result]
-        elif key == "project" and isinstance(result, list):
-            cr.project = [Project.model_validate(item) for item in result]
-        elif key == "competition" and isinstance(result, list):
-            cr.competition = [Competition.model_validate(item) for item in result]
+            return
+        model_map = {"education": Education, "work": Work, "internship": Internship, "project": Project, "competition": Competition}
+        list_map = {"education": cr.education, "work": cr.work, "internship": cr.internship, "project": cr.project, "competition": cr.competition}
+        model_cls = model_map.get(key)
+        items = list_map.get(key)
+        if model_cls is None or items is None:
+            log.warning(f"未知优化模块: {key}")
+            return
+        if isinstance(result, dict):
+            try:
+                items[idx] = model_cls.model_validate(result)
+            except (IndexError, Exception) as e:
+                log.warning(f"应用优化结果[{key}[{idx}]]失败: {e}")
+        elif isinstance(result, list) and len(result) == 1 and isinstance(result[0], dict):
+            # 兼容 LLM 偶尔返回单元素数组的情况
+            try:
+                items[idx] = model_cls.model_validate(result[0])
+            except (IndexError, Exception) as e:
+                log.warning(f"应用优化结果[{key}[{idx}]]失败(数组兼容): {e}")
+        else:
+            log.warning(f"优化结果格式异常[{key}[{idx}]]: type={type(result).__name__}")
 
     # ===== AI 对话编辑 =====