Deployment of Machine Learning Models (Udemy.com)

The initial part of the course are good. But as it progress further, especially during the deployment using training material setup based on Heroku, the quality deteriorate quickly. While the new session based on Railway is good and commendable, certain videos are theoretically applicable as not Heroku is being used. But in those old vidoes, there's no hands on demo, but basically reading the output of the code in github itself. In addition, the content of the files are a bit messy upon download, as different config files available in different folder structures. But overall the content is good as exposure and general understanding.

I found the first part of the course (Data Science) excellent and worthy of a 5-star rating. However, the deployment/MLOps section (which is the main part of the whole course) was simply impractical for me. Although, to the best of my understanding, the project content adhered to professional standards, and although I fulfil the listed course requirements, I tried to follow along, but after many hours of attempts to understand it, I became highly frustrated and eventually gave up. Whether due to the complexity of the topic or the instructional approach, I was unable to keep up. This unfortunate learning experience almost led me to abandon the idea of learning MLOps altogether. So I would highly discourage learners with little to no MLOps experience from taking this course.

This course opened my mind so much in terms of how a model can actually be used/requested in production and be able to effectivelly generate value. It made me understand how a ML model is so much more than just the model itself and the challenges that sorround translating a research model code into production code where we need to follow software engineering best practices (assure reproducibility and maintainability, testing, serving, CI/CD). This course definitelly leveled up my game and I'm just getting started!

Up to chapter 6 the course is understandable, although it requires you to look up a lot of stuff, especiall if you don't come from a software engineering background. Chapter 7 becomes too crazy for me: adding lots of dependencies at once makes it very hard to debug if something goes wrong. Also the GIT pull request moves are difficult to understand. Generally I miss a more modular structure of the course. Sections 5-7 add complex new features and functionality in parallel. Hence it is very hard to understand what each step achieves. You're just presented with an ideal solution and if it becomes too complex, the instructor says: "don't worry if this seems too complex, you rarely have to do this in practice."

The course covers the "core concepts of ML deployment", but the delivery was not good. Because of the poor delivery of content/lecture, you wouldn't get much other than overall idea of deployment. Don't expect to able to do so yourself. Showing the screen and casting it along the entire course through hundreds of lines of codes are NOT the way to teach anyone. I don't understand why it is so hard to code it along and explain your thought process, how each piece connect. At the bare minimum, instructors need to show the diagram or flow chart of how each component connect to each other. I grow frustrated as I continue toward the end of the course. From the instructors, they see the screen with all the codes in mind, all connected. For students, we see many lines of codes lumped up and we don't know what they mean, what they do, how to do it, why it does this way etc. I'm very disappointed with how the course turns out and frustrated how the course could have been done much better. It has so much potential and could have been a very good course. I wish I could give a 5 stars rating but the delivery of the content makes it impossible.

I definitely have to say WOW! This course offers detail, well - structured and comprehensive approach towards the topic about model deployment. There are many new things and resources that can be learned from this course and it is just a big step for someone who did not have a chance to work on model deployment earlier. With the way of presenting all details, lecturers showed high level of professionalism but at the same time they made this topic interesting and more desirable for further research and learning.

This course is one of the best I ever got in Udemy. It goes one step ahead by explaining fast and efficient what we should do to put our solution in production.Everyone who works as data scientist he must take that course. Strongly recommended.

This is a very valuable course for individuals who are familiar w/ Machine Learning Models but don't know how to deploy them. Highly recommended!

I am very happy with this course. I think that the content has a very high quality, super useful, the explanation is so clear! I will follow the other proposed courses. I would really like to thanks the team and the authors for this course, because I love DS, I really want to improve, and this course is a great contribution to my objective. Thanks a lot!

This course was amazing. I totally loved it. However, I must mention that the course contains a lot of Advanced material (If you have only dealt with thing machine learning algorithms and jupyter notebooks before).

Just finished section 5 and so far it's really well organised. Took me a short amount of time to get the process organised in my head.

The course I probably learned the most from. It's a challenging topic, but perfect for full-stack advocating data scientists. Being able to develop the model in a notebook is one thing, actually deploying the model to production is what makes all the impact for the business. I highly recommend this course.
Small note if you want to take this course,
Do not follow exactly what is built in the lectures, do some small variations. You will fail and struggle some time, but it's an absolutely necessary process for you to go through to really learn the skills.