What you are building in this challenge is your very own learning agent that makes you smarter. Anthropic's official documentation, your professional domain, the latest AI trends — you will learn whatever you want to master through this agent. In the AI era, there is an overwhelming amount to learn, and keeping up is exhausting. In times like these, returning to the original source is the fastest way. Instead of waiting for someone else to summarize and teach you, you will directly build a system where your agent assists your learning. In this challenge, we will build that system together over the course of four weeks. You will learn the principles through the official curriculum designed by Anthropic (Claude 101, Claude Code in Action, Prompt Engineering) and immediately apply those principles to this agent. Then, you will use that agent to continue your next steps in learning. By the end of the four weeks, you will have an agent that grows and adapts to you the more you use it!
10개 수업 학습
5회 미션 수행
라이브 5 회
성취의 증표, 수료증을 발급해요.
지식공유자와 멘토링 혜택!
같은 기수와 교류하고 함께 성장해요.
질문하고 즉시 답을 얻어요.
완주 시 쿠폰을 제공해요.
추가 자료를 통해 학습을 이어나가요.
549
Learners
51
Reviews
3
Answers
4.9
Rating
2
Courses
Simple but essential — It is simple, but it deals with the essence.
Through AX training, we will work together to create your own learning agent that adds a layer of usefulness every day, however small it may be.
01
Knowledge becomes outdated after just three months... It's overwhelming to keep up with new tools, models, and update news every day.
02
I've taken famous courses, but I don't know how to apply them to my own work and problems.
03
I barely manage to produce AI outputs, but I lack the ability to judge and verify whether those results are correct.
04
I haven't used it properly yet, but I feel overwhelmed because I think I'll fall behind if I don't start now.
So, we prepared this
The concept of this workshop is the 'Meta Loop.'
Learn the principles through the Anthropic official curriculum → Apply those principles to your own agent
→ and the goal is to create an environment where that agent helps you learn AI even better.
From prompt principles, Tool Use, Thinking & Reasoning, to Agentic Systems,
you will build the core structure of agent design step-by-step each week.
We don't just stop at ideas; we implement a single agent ourselves.
We will help you so that everyone can complete the journey to the end.
You will gain experience in directly checking and modifying whether the AI is operating according to your intentions and improving the results.
Afterward, you will be able to verify the results yourself, no matter what AI tool you use.
Over the course of 4 weeks, we will build an AI learning agent for yourself.
From learning to reflection, the agent will operate according to your specific situation and context.
You can create them in any purpose or form you want, such as CLI agents like Claude Code, background tasks, or MCP servers.
Agents you can build
(This image is a simulation, and the results each person can create may vary.)
A chat-based agent that learns through conversation
An agent that provides quizzes/problem-solving and analyzes my weaknesses
Operation Period
Number of Recruits
Challenge Structure
You will learn the principles of designing context, roles, and output formats to make the agent respond in the desired way.
🎯 What you will create in Session 1
Designing a tutor prompt that explains at my level
Creating a prompt structure that draws out the desired answers
Expand the agent's capabilities by connecting external tools such as search, document collection, and API calls.
🎯 What we will create in Session 2
Connecting official documentation search
Configuring an AI news collection workflow
Implementing Tool Use to fetch external information
Design it to reason by considering various possibilities
rather than providing a simple answer.
🎯 What we create in Session 3
Step-by-step reasoning structure for complex questions
Designing Comparative/Analytical Responses
Applying reasoning patterns to improve response quality
Design an agent structure that understands goals, plans, and executes.
🎯 What we create in Session 4
Designing goal-based execution flows
Implementing a Learning Planning Agent
Connecting multiple functions to configure them as a single system
Presenting my own AI learning agent created over the course of 4 weeks.
Share the completed version of the agent
Individual Feedback
Meta-Loop Retrospective
Post-Challenge Enhancement Roadmap
Hello, I am Lim Dong-jun, a knowledge sharer.
I am currently conducting developer training and
AI utilization training at Woowa Tech Course.
Woowahan Tech Course Frontend Coach (2022~)
Conducted numerous internal and external AI utilization training sessions and workshops (2024–)
Leading team-level AI-related assetization and educational experiments
AI tools change incredibly fast. It's common for what you learned 3 or 6 months ago to quickly become outdated. That's why I've come to focus more on principles than on the tools themselves. Prompt design, the way of working with AI, and the attitude of verifying results remain even when the tools change.
And there is one more thing I felt while experiencing many educational settings. Even a good lecture does not become yours just by watching it. It only becomes yours when you actually create and use it, even if it's something small.
Through this challenge, by building agents yourself every week,
I want to give you the experience of not just learning about AI, but making it truly yours :)
✅ Those who are tired of the daily flood of AI news and want to return to the original sources to learn deeply
✅ Those who want to master the fundamental principles and ways of thinking that remain constant even as AI tools change every month
✅ Those who want to create their own AI learning partner using prompts, even without coding experience
✅ Those who have development experience and want to design agents by directly extending them to CLI, web apps, and MCP
⚠️ Warning! This might not be the right fit for you.
For those who only want to quickly get ready-to-use prompt templates, or those who intend to only watch the recordings without participating in live sessions and missions, the utility of this challenge may not be significant.
Operating System (OS): Both Mac / Windows available
Required Accounts: Anthropic Claude (Pro or higher), GitHub
Weekly pre-lecture materials, challenge mission guides, and live session recordings provided
Regardless of coding experience, the goal of this challenge is for anyone to create their own agent.
However, since this challenge involves a process of building and giving feedback together,
active participation and lively communication are essential.
To ensure everyone can gain as much as possible,
we plan to proceed with those who truly need this time through a preliminary survey.
5월
12일
챌린지 시작일
2026년 5월 12일 오후 03:00
챌린지 종료일
2026년 6월 10일 오후 02:30
All
20 lectures ∙ (11hr 16min)
Course Materials:
My Personal Learning Agent — Creating a tool that digests Anthropic's official documentation and the latest AI news at my level.
The ability to review and improve AI outputs — gain experience in identifying where AI makes mistakes and directly refining prompts to achieve the desired results.
Practical knowledge applied to my tools using the official curriculum — you will learn and internalize concepts designed by Anthropic, such as prompt principles, Tool Use, and agent systems, by implementing them yourself.
Source-based self-learning methodology — Create a structure for learning independently through official documentation, without being dragged along by secondary summaries.
Who is this course right for?
Developers who use AI but get inconsistent results that "work sometimes and don't work other times"
Those who are exhausted from trying to keep up with the daily flood of AI news, blogs, and YouTube videos
Those who find it difficult to judge whether AI-generated code or answers are correct or incorrect
Those who gave up trying to read the official Anthropic documentation on their own
Need to know before starting?
It is sufficient if you have experience using AI chatbots such as Claude or ChatGPT.
If you have coding experience, you can create web apps, CLI, or MCP tools; even if you don't, you can complete learning tools in the form of prompt sets.
취소 및 환불 규정
챌린지는 지식공유자가 설정한 수업 최소 정원이 충족되지 않을 경우, 폐강 안내가 고지되며 결제 내역이 자동취소됩니다.
Check out other courses by the instructor!
$152.90
