基于LLM的系统设计与实现

RAG
- Overview
- 实战 *
- Pattern
  - (原理) RAG Pattern *
  - Agentic RAG
    - (原理|实战)Agentic RAG +
    - (原理|实战)Self-Reflective RAG
  - KG-RAG
    - GraphRAG +
    - RAG KG
  - Multimodal RAG
    - (Survey)多模态 RAG +
    - (实战)多模态 RAG
- phase *
  - (Phase)index
    - (原理) Index +
    - (原理|实战) Chunk +
  - (Phase)pre-retrival
  - (Phase)post-retrieval
    - (原理|实战)RAG Fusion
    - (原理|实战)RAG Rerank
- 案例
  - (原理)RAG OpenAI案例
  - (原理)RAG Baichuan案例
Agent
Agent Infra
- Agent Memory
  - Agent Memory 原理
  - Agent Memory 实践
- Memory
- Reflection
  - Reflection Agent *
- Planning
- Sandbox
  - Sandbox
- Tool use
FineTuning
- Instruct Tuning
  - (原理)Instruct Tuning
  - (Survey)Instruct Tuning
- 实践 *
- PEFT *
  - (原理)PEFT +
  - (实战)PEFT 概述
  - Soft Prompt
  - Lora
    - (原理|实战) QLoRA *
    - (实战) Lora +
- Data
  - Data Quality
    - Instruction Quality
      - (原理)LIMA, LESS
    - Instruction Diversity
      - (原理)SELF-INSTRUCT+
      - (原理|实战)Self-QA *
    - Instruction Complexity
      - (原理)Wizard
  - Task composition
    - (原理)SFT 数据组合 *
  - (原理)SFT Scaling
  - (原理)Data Selection
  - (List)SFT数据集
Prompt Engineering
Harness engineering
Langchain
Context Engineering
- (Manus)context engineering
DeepResearch
- 实现
- Overview
  - (原理&实战)技术全景
Application
- GPT-工具和应用
- 向量数据库
- NL2SQL
其他
- 自己的工作
- (原理)Agent 分类[有趣|有用]

(Manus)context engineering

总结[1]
参考

总结[1] #

参考 #

Manus 内部的 Context 工程经验（精校、高亮要点）

Manus 创始人手把手拆解：如何系统性打造 AI Agent 的上下文工程？
Context Engineering for AI Agents: Lessons from Building Manus

总结[1]
参考