惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

L
LangChain Blog
Security Latest
Security Latest
P
Proofpoint News Feed
GbyAI
GbyAI
PCI Perspectives
PCI Perspectives
博客园 - Franky
N
Netflix TechBlog - Medium
博客园_首页
WordPress大学
WordPress大学
K
Kaspersky official blog
CTFtime.org: upcoming CTF events
CTFtime.org: upcoming CTF events
Threat Intelligence Blog | Flashpoint
Threat Intelligence Blog | Flashpoint
Vercel News
Vercel News
T
Threatpost
The Hacker News
The Hacker News
H
Help Net Security
S
Securelist
Recent Announcements
Recent Announcements
腾讯CDC
T
Tailwind CSS Blog
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
cs.CL updates on arXiv.org
cs.CL updates on arXiv.org
Engineering at Meta
Engineering at Meta
C
Cisco Blogs
V
V2EX
C
Check Point Blog
S
Schneier on Security
Cyberwarzone
Cyberwarzone
C
Cybersecurity and Infrastructure Security Agency CISA
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
B
Blog RSS Feed
H
Hackread – Cybersecurity News, Data Breaches, AI and More
Jina AI
Jina AI
M
MIT News - Artificial intelligence
T
Threat Research - Cisco Blogs
博客园 - 叶小钗
A
Arctic Wolf
AWS News Blog
AWS News Blog
Latest news
Latest news
Martin Fowler
Martin Fowler
Recorded Future
Recorded Future
Last Week in AI
Last Week in AI
The GitHub Blog
The GitHub Blog
小众软件
小众软件
B
Blog
aimingoo的专栏
aimingoo的专栏
C
Cyber Attacks, Cyber Crime and Cyber Security
V
Visual Studio Blog
P
Palo Alto Networks Blog
Spread Privacy
Spread Privacy

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant Common SOC 2 Failures (Real World) Stop Vibe-Checking Your AI App: A Practical Guide to Evals How to Use SonarQube and SonarScanner Locally to Level Up Your Code Quality Your Next To-Do App Is Dead — I Replaced Mine with an OpenClaw AI Sign a Nostr event in 60 lines of Python using coincurve — no nostr-sdk, no nbxplorer, no rust toolchain ITGC Audit Explained Like You’re in Big 4 Patch Tuesday abril 2026: Microsoft parcha 163 vulnerabilidades y un zero-day en SharePoint Stop scraping everything: a better way to track competitor price changes Listing on MCPize + the Official MCP Registry while routing payments OUTSIDE the marketplace — how I kept 100% of my x402 revenue Building an AI-Powered Risk Intelligence System Using Serverless Architecture Why We Ripped Function Overloading Out of Our AI Toolchain Testing AI-Generated Code: How to Actually Know If It Works SaaS Churn Is Killing Your Business. Here Is What to Do About It (Without a Support Team) The Speed of AI Is No Longer Linear - And Self-Improving Models Are Why How to Implement RBAC for MCP Tools: A Practical Guide for Engineering Teams From Standard Quote to Persuasive Proposal: AI Automation for Arborists I built a CLI that scaffolds complete multi-tenant SaaS apps Axios CVE-2025–62718: The Silent SSRF Bug That Could Be Hiding in Your Node.js App Right Now The dashboard that ended our friendship Data Pipelines Explained Simply (and How to Build Them with Python) The Hidden Cost of AI Systems Nobody Talks About. undefined vs undeclared, and how typeof behaves Switching from file-based jobs to NATS/Kafka in Rust without changing code io_uring Adventures: Rust Servers That Love Syscalls Why Agentic AI is Killing the Traditional Database The POUR principles of web accessibility for developers and designers Quantum Neural Network 3D — A Deep Dive into Interactive WebGL Visualization How To Install Caveman In Codex On macOS And Windows Automation Pipeline Reliability: Why Your Workflow Breaks When Nobody Is Watching I Built an 'Open World' AI Coding Agent — It Works From ANY Folder From Freelancing to Product: A Tech Service Company's SaaS Transformation China's AI Giants: Adding Tencent Hunyuan & ByteDance Doubao to AI University (74 Providers) On the Vibe Coders and Their Lies clerk: Auto-Summarize Your Claude Code Sessions AI Weekly — 2026/04/10–04/17 | The Model Lockdown Is Here, but the Toolchain Is the Real Battleground AI 週報 — 2026/04/10–2026/04/17 模型封鎖潮來了,但工具鏈才是真戰場 Maybe this is how Open-Source apps are born... 🚀 Fine-Tune LLMs with LoRA and QLoRA: 2026 Guide tRPC v11 + Next.js App Router: End-to-End Type Safety Without the Boilerplate ShadCN UI in 2026: Why I Stopped Installing Component Libraries and Started Owning My Components SaaS Billing in React Server Components: Stripe + Supabase Without a Single `useEffect` Join our DEV Weekend Challenge — $1,000 in Prizes Across TEN winners! Submissions Due April 20 at 6:59 AM UTC. Implementing FSRS Spaced Repetition in Flutter + Supabase — Adding Memory Science to an AI Learning App "I Texted My Localhost From the Train — Claude Code Fixed the Bug Before I Got Home" I Built a Sales Prep AI and It Went Deeper Than Expected Design to Code #2: One JSON, Eleven Outputs Solving the 100M-Row Problem: A Summary Table Pattern for High-Volume Push Notification Logs Flutter Web With Wasm: What Actually Changes For Developers I Built 50 Royalty-Free Soundtracks for My Side Project in a Weekend Using AI Music Generation The Vibe Coding Security Checklist: 7 Things to Check Before You Ship Stop Letting Googlebot Guess Fix Your React App's SEO Right Desconstruindo o Streaming do LinkedIn: Como Criar um Engine de Extração de Vídeo de Alta Performance com HLS e FFmpeg (EDA Part-1) EDA (Exploratory Data Analysis) Explained With Real Life — Why Looking at Your Data Is the Most Important Step in Machine Learning Brand Relationship Management at Scale: Our 4-Touch Outreach System for 200+ Brands Why String.fromEnvironment() Might Return an Empty String in Dart JGuardrails 1.0.0 — Hardening Java LLM Apps Against Jailbreaks, Toxicity, and Prompt Injection Plan and Schedule a Full Week of Threads Content From One Claude Conversation Coding Cat Oran Ep3, Five Tables Changed Everything BFF模式详解:构建前后端协同的中间层 I'm done watching freelancers get buried by 200 proposals. So I'm building the alternative. This is my first post BFS Algorithm in Java Step by Step Tutorial with Examples Tracking LLM Pricing Monthly: An Open Dataset for 22 AI Models How We Measure Content ROI on a Comparison Site: Revenue Attribution Without Perfect Data Introducing Nova AI Ops: The AI-Native Operating System for SRE Teams I built a free desktop video downloader for Windows — Grabbit How Talkie OCR Helps Vision-Impaired & Dyslexic Users Read the World Around Them VRCFaceTracking安装和iPhone面捕配置教程,有bug Even CrowdStrike Can't See Your Agents The Automation Gold Rush: What n8n Workflows and Claude Are Opening Up for Developers Right Now
RAG 시스템 실전 구축 (v3)
matias yoon · 2026-05-24 · via DEV Community

matias yoon

RAG 시스템 실전 구축 (v3)

개요

이 가이드는 실전에서 RAG(Retrieval-Augmented Generation) 시스템을 구축하는 데 필요한 모든 것을 다룹니다. ML 엔지니어와 백엔드 개발자들이 데이터를 기반으로 의미 있는 질문에 답변할 수 있는 시스템을 구축할 때 실용적인 코드와 접근 방식을 제공합니다.

1. RAG 기본 개념

RAG는 검색 기반 생성 시스템으로, 두 가지 주요 단계를 거칩니다:

검색 (Retrieval)

문서에서 관련 정보를 찾는 과정

보강 (Augmentation)

검색된 정보를 프롬프트에 추가하여 LLM이 답변 생성에 활용

생성 (Generation)

LLM이 검색된 정보와 원본 질문을 기반으로 답변 생성

질문 → 검색 → 보강 → 생성 → 답변

Enter fullscreen mode Exit fullscreen mode

2. 청킹 전략

2.1 의미적 청킹 (Semantic Chunking)

from sentence_transformers import SentenceTransformer
import numpy as np
from sklearn.cluster import KMeans

def semantic_chunking(text, model, threshold=0.7):
    # 텍스트를 문장 단위로 분할
    sentences = text.split('. ')
    sentence_embeddings = model.encode(sentences)

    # 클러스터링을 통한 의미적 청킹
    kmeans = KMeans(n_clusters=min(len(sentences), 10))
    kmeans.fit(sentence_embeddings)

    # 같은 클러스터 내 문장들을 하나의 청크로 결합
    chunks = []
    for i in range(len(sentences)):
        cluster = kmeans.labels_[i]
        if cluster not in [c[0] for c in chunks]:
            chunks.append((cluster, [sentences[i]]))
        else:
            for j, (c_cluster, c_sentences) in enumerate(chunks):
                if c_cluster == cluster:
                    chunks[j] = (c_cluster, c_sentences + [sentences[i]])
                    break

    return [chunk[1] for chunk in chunks]

# 사용 예시
model = SentenceTransformer('all-MiniLM-L6-v2')
text = "AI 기술은 빠르게 발전하고 있습니다. 인공지능은 많은 분야에서 활용됩니다. 특히 자연어 처리 분야에서는 큰 진전이 있습니다."
chunks = semantic_chunking(text, model)

Enter fullscreen mode Exit fullscreen mode

2.2 재귀적 청킹 (Recursive Chunking)

import re

def recursive_chunking(text, chunk_size=512):
    # 문단 단위 청킹
    paragraphs = text.split('\n\n')
    chunks = []

    current_chunk = ""
    for para in paragraphs:
        if len(current_chunk) + len(para) < chunk_size:
            current_chunk += para + "\n\n"
        else:
            if current_chunk:
                chunks.append(current_chunk.strip())
            current_chunk = para + "\n\n"

    if current_chunk:
        chunks.append(current_chunk.strip())

    return chunks

Enter fullscreen mode Exit fullscreen mode

2.3 에이전트 기반 청킹 (Agentic Chunking)

class AgenticChunker:
    def __init__(self, model_name="gpt-3.5-turbo"):
        self.model = model_name

    def chunk_with_context(self, text, context_length=1000):
        # 문맥을 고려한 청킹
        sentences = re.split(r'[.!?]+', text)
        chunks = []
        current_chunk = ""

        for sentence in sentences:
            if len(current_chunk) + len(sentence) < context_length:
                current_chunk += sentence + ". "
            else:
                if current_chunk.strip():
                    chunks.append(current_chunk.strip())
                current_chunk = sentence + ". "

        if current_chunk.strip():
            chunks.append(current_chunk.strip())

        return chunks

Enter fullscreen mode Exit fullscreen mode

3. 임베딩 모델 선택

3.1 비교 테이블

모델 성능 속도 메모리 용도
all-MiniLM-L6-v2 86.4 빠름 100MB 일반
all-mpnet-base-v2 88.2 중간 400MB 정확도
BGE-M3 90.1 중간 500MB 다국어
Sentence-BERT 89.8 느림 700MB 높은 정확도

3.2 성능 비교 코드

from sentence_transformers import SentenceTransformer
from sklearn.metrics.pairwise import cosine_similarity
import time

def evaluate_embedding_models():
    models = {
        'all-MiniLM-L6-v2': SentenceTransformer('all-MiniLM-L6-v2'),
        'all-mpnet-base-v2': SentenceTransformer('all-mpnet-base-v2'),
        'BGE-M3': SentenceTransformer('BAAI/bge-m3')
    }

    test_sentences = [
        "Machine learning is a subset of artificial intelligence.",
        "Deep learning uses neural networks with multiple layers.",
        "Natural language processing deals with human language understanding."
    ]

    results = {}
    for name, model in models.items():
        start_time = time.time()
        embeddings = model.encode(test_sentences)
        end_time = time.time()

        # 유사도 계산
        similarity = cosine_similarity([embeddings[0]], [embeddings[1]])[0][0]

        results[name] = {
            'time': end_time - start_time,
            'similarity': similarity,
            'memory': model._modules['0'].state_dict()['weight'].numel() * 4 / (1024**2)  # MB
        }

    return results

# 성능 평가
perf_results = evaluate_embedding_models()
for model, metrics in perf_results.items():
    print(f"{model}: Time={metrics['time']:.3f}s, Similarity={metrics['similarity']:.3f}")

Enter fullscreen mode Exit fullscreen mode

4. 벡터 데이터베이스 비교

4.1 Chroma (가볍고 빠름)

import chromadb
from chromadb import Client

class ChromaVectorStore:
    def __init__(self, collection_name="documents"):
        self.client = Client()
        self.collection = self.client.get_or_create_collection(collection_name)

    def add_documents(self, documents, embeddings, metadatas=None):
        self.collection.add(
            documents=documents,
            embeddings=embeddings,
            metadatas=metadatas
        )

    def search(self, query_embedding, n_results=5):
        results = self.collection.query(
            query_embeddings=[query_embedding],
            n_results=n_results
        )
        return results

# 사용 예시
chroma_store = ChromaVectorStore()
# embeddings = model.encode(documents)
# chroma_store.add_documents(documents, embeddings)

Enter fullscreen mode Exit fullscreen mode

4.2 Qdrant (고성능)

from qdrant_client import QdrantClient
from qdrant_client.models import Filter, FieldCondition, MatchValue

class QdrantVectorStore:
    def __init__(self, host="localhost", port=6333, collection_name="documents"):
        self.client = QdrantClient(host=host, port=port)
        self.collection_name = collection_name

        # 컬렉션 생성
        self.client.recreate_collection(
            collection_name=self.collection_name,
            vectors_config={"size": 384, "distance": "Cosine"}
        )

    def add_documents(self, documents, embeddings, metadatas=None):
        points = [
            {
                "id": i,
                "vector": embedding,
                "payload": {"text": doc, "metadata": metadata if metadata else {}}
            }
            for i, (doc, embedding, metadata) in enumerate(zip(documents, embeddings, metadatas or [None]*len(documents)))
        ]
        self.client.upsert(
            collection_name=self.collection_name,
            points=points
        )

    def search(self, query_embedding, n_results=5):
        results = self.client.search(
            collection_name=self.collection_name,
            query_vector=query_embedding,
            limit=n_results
        )
        return results

# 사용 예시
qdrant_store = QdrantVectorStore()
# qdrant_store.add_documents(documents, embeddings)

Enter fullscreen mode Exit fullscreen mode

4.3 pgvector (SQL 기반)


python
import psycopg2
from psycopg2.extras import Json
import numpy as np

class PGVectorStore:
    def __init__(self, connection_string):
        self.conn = psycopg2.connect(connection_string)
        self.create_table()

    def create_table(self):
        with self.conn.cursor() as cur:
            cur.execute("""
                CREATE TABLE IF NOT EXISTS document_embeddings (
                    id SERIAL PRIMARY KEY,
                    document TEXT,
                    embedding VECTOR(384),
                    metadata JSONB,
                    created_at TIMESTAMP DEFAULT NOW()
                )
            """)
            cur.execute("CREATE INDEX IF NOT EXISTS idx_embedding ON document_embeddings USING ivfflat (embedding vector_cosine_ops)")
        self.conn.commit()

    def add_documents(self,

---

📥 **Get the full guide on Gumroad**: https://gumroad.com/l/auto ($7)

Enter fullscreen mode Exit fullscreen mode