self-rag

A complete collection of RAG interview questions, answers (1509 questions & 52 RAG types), system design scenarios, architecture patterns, and production-ready concepts.

按问题复杂度自动分流的电商客服 RAG：简单问题直接回答；复杂问题经过多跳检索并自我纠错；查单、退改通过工具调用完成，且高风险写操作必须先经人工确认。基于 FastAPI 和 LangGraph，混合检索并附带引用，内置离线评测与 Prometheus 指标。涵盖简历写法和项目面试拷打

生产级 3GPP 5G 规范 RAG Agent：自然语言提问，回答带段落级原文引用 + 严格 grounding，覆盖 Rel-18/19 全部 TS。

Evidence-synthesis RAG assistant for TCM practitioners — hybrid vector + knowledge graph retrieval over 17 classical texts, with query classification, self-critique verification, and blind A/B arena evaluation.

Training code for advanced RAG techniques - Adaptive-RAG, Corrective RAG, RQ-RAG, Self-RAG, Agentic RAG, and ReZero. Reproduces paper methodologies to fine-tune LLMs via SFT and GRPO for adaptive retrieval, corrective evaluation, query refinement, self-reflection, and agentic search behaviors.

ATLAS — industrial Agentic RAG platform (LangGraph): routing, query decomposition, Corrective RAG, Self-RAG reflection. FastAPI + SSE, pluggable vector stores, auth, rate limiting, tracing, Docker, CI. Academic webinar kit on archive/academic-labs branch.

Self-Reflective Question Answering for Biomedical Reasoning. GRPO fine-tuning via QLoRA & Unsloth with rewards for correctness, relevance, groundness, utility & XML structure. Structured think → answer → self-reflection with context grading, relevance assessment & groundness evaluation. DeepEval LLM-as-a-Judge (GEval, Faithfulness, Relevancy).

RAG enhances LLMs by retrieving relevant external knowledge before generating responses, improving accuracy and reducing hallucinations.

Self-correcting agentic RAG with LangGraph — combines Adaptive, Corrective, and Self-RAG. Routes queries, grades retrieved documents, and checks answers for hallucinations. Runs fully local on Ollama.

Learn Retrieval-Augmented Generation (RAG) from scratch by manually building chunking, embeddings, retrieval, reranking, answerability, local generation, Graph-RAG, and Self-RAG.

基于 LangGraph 的 Agentic RAG 知识库助手，支持混合检索、Self-RAG、引用溯源和 RAGAS 自动评估。

Agentic RAG system with five workflow modes from Naive to Deep, routing each question to the cheapest path that answers it well, fusing dense and BM25 retrieval through Reciprocal Rank Fusion with CrossEncoder reranking, verifying groundedness and citations, and gating risky answers behind a human-in-the-loop approval queue.

Hands-on implementation of modern Retrieval-Augmented Generation (RAG) systems using LangChain, LangGraph, vector databases, RAGAS, and Agentic AI workflows.

LangGraph implementation of Self-RAG, based on “Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection” (Asai et al., ICLR 2024), implementing adaptive retrieval, evidence evaluation, generation, and self-critique inspired by the paper’s reflection-token architecture.

Self-RAG chatbot with LangGraph that retrieves from internal PDFs, verifies grounding (IsSUP), checks usefulness (IsUSE), and iteratively rewrites queries to improve answer quality.

企业级 Agentic RAG 技术文档问答系统

Production-ready Retrieval-Augmented Generation (RAG) system with hybrid retrieval, Self-RAG agent workflows, cross-encoder reranking, and comprehensive benchmarking.

Backend Of STELLA that run SELF Retrieval-Augmented Generation [Senior Project]

A self-correcting RAG agent in LangGraph — routes questions, grades its own retrieval, falls back to web search, and refuses answers that aren't grounded in its sources.

Self-RAG: a self-reflective RAG system built from scratch in LangGraph. Grades its own retrieved documents and generated answers (relevance, grounding, usefulness), then self-corrects via bounded regenerate/re-retrieve loops. MAANG-standard structure with config, custom exceptions, logging, and tests.