Inflearn brand logo image
Inflearn brand logo image
Inflearn brand logo image
BEST
AI Development

/

AI Agent Development

Developing LLM Applications Using RAG (feat. LangChain)

RAG. Learn from Silicon Valley GenAI Hackathon Winner. Packed with real-world know-how.

(4.9) 243 reviews

2,370 learners

  • jasonkang
이론 실습 모두
NLP
LLM
RAG
LangChain
vector-database
openAI API

Reviews from Early Learners

What you will learn!

  • LangChain

  • Large Language Model

  • Vector Database

  • Retrieval Augmented Generation(RAG)


RAG Crafted by a Silicon Valley GenAI Hackathon Winner

  • Data Preprocessing and Efficient Retrieval: Learn data preprocessing techniques necessary for RAG configuration and methods to maximize search efficiency using keywords.

  • Efficient Prompt Writing Methods: With improved LLM performance, you can now achieve good results even when writing prompts in Korean. Learn how to write Korean prompts using LangChain's PromptTemplate.

  • LLM Performance Evaluation and Service Optimization: Learn how to systematically measure and optimize model performance, reliability, and accuracy through LLM evaluation after service deployment.

RAG? Retrieval-Augmented Generation?

RAG is Retrieval Augmented Generation, a technology that enhances the performance of Large Language Models (LLMs). LLMs have excellent language understanding and generation capabilities by learning from vast amounts of text data, but they have limitations such as bias and factual errors. RAG can complement these limitations through real-time information retrieval and improve accuracy and reliability.

Features of This Course

📌 This contains the know-how learned through hands-on experience while developing, deploying, and operating LLM Applications in real-world business environments

📌 10% theory, 90% practice. Only essential theory is explained briefly, and all lectures consist of live coding

📌 I intentionally did not edit the errors. You can learn how to debug while developing LLM services.

📌 100% lecture question resolution! We solve difficulties encountered in lecture content or real-world work together through Q&A

I recommend this for people like this

I don't know where to start.
Developers/development teams who want to create services using LLM, but feel
overwhelmed about where to begin

What is RAG?
I'm curious about what RAG is and why it's important. For those who want to understand the latest technology and use it to develop their own AI applications.

What should we do next?
Developers/Development teams
who need to solve
Hallucination issues during LLM Application development

After taking the course

  • Data Preprocessing and Keyword Utilization: You can learn data preprocessing techniques necessary for RAG configuration and methods to maximize search efficiency by utilizing keywords.

  • Model Performance Evaluation: Through LLM evaluation, you'll learn how to systematically measure and optimize a model's performance, reliability, and accuracy. You'll learn how to improve model quality through evaluation results.

  • Service Deployment and Maintenance: You'll learn how to deploy and maintain applications using tools like Streamlit, and acquire skills that can be immediately applied in practical work.

  • Solving Hallucination Problems: You'll learn techniques to minimize inaccurate information generated by LLM models and provide more reliable information.

  • Understanding and Applying the Latest AI Technologies: You can understand cutting-edge AI technologies like RAG and use them to develop your own AI applications

You'll learn this content.

LLM Response Streaming

If users keep seeing
a loading screen while the LLM is generating responses, it will feel like the service is slow. Learn how to
improve user experience through streaming.

Providing Sources for Answers

Hallucination, the biggest problem in LLM services.
Learn how to improve the reliability of answers by providing users with the sources of answers
while generating responses

LLM Evaluation Using LangSmith

During service operation, the Knowledge Base also changes,
and prompts need to be modified as well. Every time there's an update,
developers can't test each one individually.
For stable service operation, learn how to verify model accuracy using LangSmith

LangChain Expression Language (LCEL)

Did you know that LangChain can be used by connecting various chains together? Using LCEL syntax,
you'll learn how to implement and connect chains with different functionalities

Vector Database(Chroma, Pinecone)

Using LangChain with Vector Databases like Chroma and Pinecone
to store data and retrieve relevant documents through similarity search

The person who created this course



Pre-enrollment Reference Information

Practice Environment

  • The course is explained based on MacOS.

    • If Python runs on Windows and Linux environments, you can follow along.


Learning Materials

  • Source Code GitHub Repository (Jupyter Notebook, Streamlit)

  • Supplementary GitBook for Additional Explanations

Prerequisites and Important Notes

  • Python Basic Syntax

  • Anyone who has used ChatGPT even once will be able to understand this easily.

  • This would be most helpful for those who have experienced difficulties while developing LLM Applications.

Recommended for
these people

Who is this course right for?

  • A developer looking to build an LLM service

  • Developer with LLM Application development experience

  • Developer struggling with RAG setup

Need to know before starting?

  • Python

Hello
This is

11,698

Learners

588

Reviews

342

Answers

4.9

Rating

9

Courses

Curriculum

All

25 lectures ∙ (3hr 36min)

Course Materials:

Lecture resources
Published: 
Last updated: 

Reviews

All

243 reviews

4.9

243 reviews

  • Hyuntak Lee님의 프로필 이미지
    Hyuntak Lee

    Reviews 1

    Average Rating 5.0

    5

    100% enrolled

    랭체인에 이런 명강의가 없습니다! 제가 랭체인 혼자 한 번 파보겠다고 시중에 나온 랭체인 관련 도서는 싹다 구입하고, 강의도 엄청 구매하고 했는데,, 랭체인의 버전업 때문에 무용지물이 된 게 한둘이 아니고, 초심자 입장에서 이해되게 설명하는게 없었어요. 그런데 강병진님 강의는 버전도 다 최신인데다가 듣다보면 로직이 하나하나 다 이해가 돼요. 랭체인을 배우고 싶은 모두에게 강추하고 싶은 강의입니다

    • 강병진
      Instructor

      감사합니다! 강의의 장점은 코드가 Deprecate되면 추가로 업데이트가 가능한 점이라고 생각합니다. 혹시나 랭체인에서 업데이트가 발생하면 저도 강의를 업데이트하고 말씀드릴게요~

  • swgoodcode님의 프로필 이미지
    swgoodcode

    Reviews 9

    Average Rating 5.0

    5

    95% enrolled

    느낀점 1) 내가 개인 llm 서비스를 만들었다고? 실화?…쉽다… 2) 강사님은 이 많은 기술 어떻게 배운거야…내가 계속 나아갈 수 있을까? 좌절… 하지만 문득 아이디어가 떠오르더군요… 강사님을 독촉해서 다음 강의를 만들게하자… 기다리겠습니다! 빨리 다음 강의 부탁드려요!! 실무에서 인터넷이 불가능한 환경이라 라마3로 하는 방법이 궁금합니다. 그리고 랭스미스에 대하여 잘 다루고 싶어졌습니다.

    • 강병진
      Instructor

      "강의를 듣고 서비스를 구현할 수 있는 수준" 이라는 인정을 받은 것 같습니다. 정말 감사합니다. 질문으로 올려주신 내용에 답변 드린 것처럼 LLM Evaluation은 별도의 강의를 촬영할 예정입니다. 강의가 나오면 제일 먼저 알려드릴게요!

  • host0806님의 프로필 이미지
    host0806

    Reviews 1

    Average Rating 5.0

    5

    95% enrolled

    간단한 챗봇을 만들어보려고 며칠을 챗지피티와 씨름했던 시간이 무색하게 정말 이해하기 쉽게 개발 과정을 배울 수 있었습니다. 혼자 공부할 때와는 다르게 어떤게 어디에 필요한지 알 수 있어 너무 좋았습니다. 윈도우 사용자도 크게 어렵지 않게 따라갈 수 있었습니다. 추천합니다

    • 강병진
      Instructor

      따흑 감사합니다 ☺️ 처음 해보시는 분들도 쉽게 따라가실 수 있도록 기획했는데 잘 받아주신 것 같아요! 저는 앞으로 LLM 관련 업무를 계속 할거라 도움 드릴 수 있는 자료들을 많이 공유드려보겠습니다!

  • JAY probio님의 프로필 이미지
    JAY probio

    Reviews 7

    Average Rating 5.0

    5

    69% enrolled

    짧고 강력한 강의입니다. 핵심만 쏙쏙 뽑아서 전달하는 강의입니다. 다른 강의 들어보니까 돌려돌려 설명해서 더 어렵기만하고 이해도 안되고 ㅠㅠㅠㅠㅠㅠ 다른 강의도 많이 만들어주세요

    • Dominus Mr.님의 프로필 이미지
      Dominus Mr.

      Reviews 6

      Average Rating 5.0

      Edited

      5

      100% enrolled

      강의가 굉장히 재밌고 역동적입니다. 그리고 알고 싶고 궁굼했던 내용이 대부분이라 짱짱입니다 인프런에서 결제 꽤 많이 했는데, 몇번씩 돌려 보고 또보고 하는 건 처음 인 것 같습니다. 확실히 개념이 잡혀 갑니다. 진짜 감사합니다. 진짜 추천합니다.

      $51.70

      jasonkang's other courses

      Check out other courses by the instructor!

      Similar courses

      Explore other courses in the same field!