Thực hành kỹ sư máy học

Good points 1. You can learn practical approaches to fairly deep and realistic topics in a short period of time. 2. You can see useful tools such as wandb, wit, and implementation processes such as docker, kubernetes, kubeflow, mlflow, and bentoML. 3. You can obtain practical/practical approaches or know-how. Bad points 1. It is not a lecture just for 'Inflearn lectures', but it seems to have been recorded and uploaded as is from other talks? or lectures? => There are many distracting parts, and the lecture recording quality is a bit disappointing compared to other lectures for the same price. I purchased and listened to quite a few lectures from Inflearn, and watched lectures on other platforms, but if you only look at the lecture quality, it would be difficult to get a good evaluation. 2. Continuing from 1, the unrefined explanations/personal opinions/lecture materials are disappointing. It seems like something that could happen because it wasn't filmed just for 'Inflearn'. (I think Inflearn doesn't do any pre-review of the lecture content at all.) 3. Since you have to leave a course review to receive the lecture materials, it's difficult to follow along while watching the video. Of course, it might be different if you write a course review before completing the course, but I personally don't want to leave a useless review, so I plan to leave a review after completing the course and watch the lecture materials while listening to it repeatedly. Overall As it was announced that the lecture target is 'intermediate or higher', I don't recommend it if you don't have some experience or have heard of it. However, it's a good opportunity to quickly look through it while you have experience and see the approach from a practitioner's perspective. Also, if you're someone who wants to study machine learning/deep learning models diligently and acquire MLOps skills, it's not bad to listen to it once. If you take this course and also complete the MLOps course that recently opened on Coursera, it seems like a good combination.

There aren't many Korean materials on MLOps, so I appreciate you for giving a good lecture so that others can easily access it. However, if you had paid more attention to the editing part, it would have been much better (I can't help but feel like you just uploaded a recorded lecture from somewhere else...). Overall, it seems like a good lecture to take a quick look at.

It was a good lecture, but the process of setting up the environment was an old version, so it took a long time to solve the problem. I think the lecture needs to be updated or maintained. Of course, the lecture itself is really good. Thank you.

In the ML field, the most time-consuming and costly thing is not model development, but data preparation (preprocessing, transformation, etc.) and deploying and continuously maintaining the model for operation after learning the ML model. The tool that supports this is MLOps, and this course is very helpful for practical work in a situation where there is not much information about MLOps.

It's good that I got to know many useful tools while taking the lecture. However, it's a bit disappointing that the lecture is not targeted at Infraon, but is copied from another seminar.