# aws-ai-ml-workshop-kr: a Korean-language AWS AI/ML workshop repository

> This is a set of hands-on lab notebooks and workshop materials for AWS AI and ML services, written in Korean. It is a teaching repository, not a library, and that distinction decides whether you should clone it.

**aws-samples/aws-ai-ml-workshop-kr** — A collection of localized (Korean) AWS AI/ML workshop materials for hands-on labs.

- Repository: https://github.com/aws-samples/aws-ai-ml-workshop-kr
- Stars: 308 · Forks: 167
- Language: Jupyter Notebook
- License: MIT-0
- Published: 2026-09-10 · Updated: 2026-09-10 · Language: en
- Canonical page: https://hysenlabs.com/projects/aws-samples-aws-ai-ml-workshop-kr

## What aws-ai-ml-workshop-kr is, and the problem it removes

Anyone who has taught an AWS machine learning class knows the preparation cost is not the slides. It is the notebooks: a training job that actually completes, an endpoint that actually responds, a permission set that does not fail halfway through a lab. This repository collects that preparation in Korean. The README describes it as "AWS AIML 코드 및 워크샵 예제", code and workshop examples, and organises the top level into three named areas: genai/ for generative AI, sagemaker/ for the SageMaker platform, and neuron/ for AWS Neuron on Inferentia, Inferentia2 and Trainium hardware. A fourth directory, z-archive/, sits alongside them.

The audience is narrow and identifiable: instructors and students running AWS AI/ML workshops in Korean, plus engineers who read Korean more comfortably than English and want a worked example rather than API documentation. If your team works in English, the value drops sharply, because the explanatory prose in the notebooks is the part you would be translating. The code cells themselves are Python and would survive translation, but the surrounding narration is the reason the repository exists in this form.

## How the repository is organised, and what that implies about maintenance

The layout is a three-way split by AWS service family rather than by difficulty or by lab length. Under genai/ the README points to a nested path, genai/aws-gen-ai-kr/, which itself splits into 20_applications, 30_fine_tune and 40_inference. Those numeric prefixes are the only ordering signal in the repository. Under sagemaker/ there is a Readme.md plus 01-sagemaker-101 for an introductory training lab and hyperpod for model training on SageMaker HyperPod. Under neuron/ there is another Readme.md.

That structure tells you something practical: the entry point is not the root README, it is the README inside each subdirectory. The root file is a directory map with links, and it explicitly says to consult the per-directory Readme files. If you clone the repository and start reading from the top, you will get a table of contents, not instructions.

The repository carries no releases. There are no version tags to pin, so the only stable reference you have is a commit hash on master. The last push to master was on 2026-08-04, which is recent enough that the notebooks have not gone stale, but the absence of releases means you cannot ask "which version of this workshop did we teach last quarter?" without recording the commit yourself. For a training repository that is a mild annoyance. For anything you intended to depend on, it would be disqualifying.

## Installing nothing: how to get the notebooks and run the first lab

There is no package to install. The README does not document a pip install, an npm package, a Docker image or a CLI. The distribution mechanism is git clone plus the AWS console and, for the SageMaker material, a SageMaker notebook instance or SageMaker Studio environment where the notebooks can execute. The README does not spell out that environment setup, so treat the following as the mechanical part only.

Clone the repository and look at the three top-level areas before choosing one:

```bash
git clone https://github.com/aws-samples/aws-ai-ml-workshop-kr.git
cd aws-ai-ml-workshop-kr
ls genai sagemaker neuron
```

You should see the directories named in the root README, including genai/aws-gen-ai-kr/ with its 20_applications, 30_fine_tune and 40_inference subdirectories. From there, read the Readme.md inside the area you picked, because that file is where the README directs you for actual guidance.

If your chosen lab is SageMaker-based, the practical path is to upload the notebook into a SageMaker environment and run it there rather than locally, since the notebooks assume AWS credentials and service access. The repository does not include an environment.yml, a requirements.txt at the root, or a documented IAM policy, so role and permission setup is on you. That is the single biggest gap for a first-time user: the notebooks tell you what to call, not what your execution role must be allowed to call.

## The licence is stated two different ways, and that matters

The repository metadata says MIT-0. The README's own License section says the library is licensed under the Apache 2.0 License and points to the LICENSE file. Both statements appear in the repository, and they are not the same licence. MIT-0 removes the attribution requirement entirely; Apache 2.0 keeps attribution and adds an explicit patent grant.

For a workshop repository the practical difference is small, because you are running notebooks rather than redistributing a binary. It becomes real if you copy notebook code into a product, or if you republish the notebooks with your own branding. The LICENSE file at the repository root is the authoritative artefact, and the badge in the README says MIT while the prose says Apache 2.0. Read the file, not the badge, and not this article. I am not giving legal advice; I am pointing out that the repository contradicts itself and you should resolve that before reuse.

## Where this repository is the wrong tool

Three cases. First, if you need a maintained library with a version number and a changelog, this is not it. There are no releases, no semantic versioning, and the contribution section of the README says the maintainers are "still working on the best mechanism to take in examples from external sources" and asks for patience if pull requests take longer than expected or are closed. That is an honest statement of a slow intake process, and it means you should not plan to upstream a fix and rely on it landing.

Second, if you are teaching in English, the Korean narration is dead weight and you would be better served by the English-language AWS workshop material that this repository is a localisation of. The code patterns are not novel; the localisation is the product.

Third, if you need cost predictability. The notebooks target real AWS services including SageMaker training jobs, endpoints and Neuron instances. The README does not document cost estimates, cleanup steps or teardown procedures, and it does not document rollback. A workshop that leaves a training job or an endpoint running is a bill, and nothing in the repository tells you how to avoid that. Budget for cleanup as your own responsibility.

## Alternatives, and the actual difference in approach

The closest alternative is the upstream English AWS workshop and sample material published under the aws-samples organisation. The difference is not technical coverage, it is language and framing: the upstream material assumes an English-reading audience and does not carry the Korean explanatory text, while this repository is explicitly a localised collection. If your students read English, the upstream material gives you a wider surface. If they do not, this repository removes a translation step that would otherwise sit between your students and the lab.

A second alternative is building the labs yourself against the AWS service documentation. That gives you full control over the IAM policies, the cleanup steps and the cost envelope, all of which this repository leaves to you anyway. The trade is time: you would be writing the notebook cells that this repository already contains. For a one-off internal session, writing your own is often cheaper than adapting someone else's. For a recurring curriculum, starting from existing notebooks and fixing the gaps is usually faster.

## Maintenance cost and what to check before you teach from it

The last push was on 2026-08-04, so the notebooks are current. There is no release cadence to follow and no upgrade path to plan, because there is nothing to upgrade: you clone, you teach, and you re-clone when you want newer content. The cost you carry is drift. AWS service consoles and SDK defaults change, and a notebook that ran cleanly can fail on a renamed parameter or a changed default instance type. With no releases and no changelog, you have no signal that a given notebook has been re-verified.

Before a session, run the notebook end to end in the account you will teach from, not a scratch account, because quota availability differs. Check the per-directory Readme.md files for anything the root README omits. Decide explicitly whether z-archive/ is in scope; the name suggests it is not, and the root README does not list it among the three categories it presents as representative. And write down the commit hash you taught from, since that is the only version identifier this repository offers.

## Conclusion

Adopt it if you run Korean-language AWS training and want notebooks that already walk through SageMaker, generative AI applications and Neuron targets, accepting that you must supply your own AWS account, quotas and cleanup. Do not adopt it as a dependency, as a Python package, or as a source of production code, and do not treat the z-archive directory as current. Before you commit to it, open the Readme.md in each of genai/, sagemaker/ and neuron/ and confirm the notebooks still match the service console you are teaching against, because the repository ships no releases and no version tags to pin.

## FAQ

### How do I install aws-ai-ml-workshop-kr?

There is nothing to install. The README documents no package, Docker image or CLI, so the practical route is to clone the repository with git and run the notebooks inside an AWS environment such as a SageMaker notebook instance or SageMaker Studio, where they have the credentials and service access they assume.

### What licence does aws-ai-ml-workshop-kr use?

The repository metadata says MIT-0, but the README's License section says the library is licensed under the Apache 2.0 License and points to the LICENSE file. The two statements conflict, so read the LICENSE file at the repository root before reusing notebook code.

### What are the main areas of aws-ai-ml-workshop-kr?

The README presents three representative categories: genai/ for generative AI examples and service applications, sagemaker/ for SageMaker training and HyperPod material, and neuron/ for AWS Neuron on Inferentia, Inferentia2 and Trainium. A fourth directory, z-archive/, also exists at the top level but is not listed among the three.

### Is aws-ai-ml-workshop-kr available in English?

The repository describes itself as a collection of localized Korean workshop materials, and the explanatory narration in the notebooks is in Korean. The README does not document an English edition of this repository.

## Sources

- [aws-samples/aws-ai-ml-workshop-kr on GitHub](https://github.com/aws-samples/aws-ai-ml-workshop-kr)
- [Issues](https://github.com/aws-samples/aws-ai-ml-workshop-kr/issues)
- [License: MIT-0](https://github.com/aws-samples/aws-ai-ml-workshop-kr/blob/master/LICENSE)
- [README](https://github.com/aws-samples/aws-ai-ml-workshop-kr/blob/master/README.md)

---

Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/aws-samples-aws-ai-ml-workshop-kr