Hello! This is Akhil Theerthala. You might have seen my recent article series on MLOps. In that series, we dive into the lifecycle of a Machine learning project. We take a look at different phases in the lifecycle. And for each stage, we look into the common challenges and standard practices.
I have mentioned that that series of articles is based on the MLOps specialization from Deeplearning.ai. Specifically, it is based on the course, Introduction to Machine learning engineering in production, taught by the legend Andrew Ng. In this article, I will be giving an in-depth review of the course, what I liked about it, what I didn't like and so on.
Overview of this course
This course has a simple structure. It first gives the structure of a machine learning project lifecycle. And slowly covers individual phases. The discussion about each step covers the standard practices and challenges faced in the industry.
The week-wise split is given below,
- Machine learning project lifecycle
- Overview of the deployment phase.
- Discussions about the Modelling phase of the project.
- Slight overview of the Data phase of the project.
- Continuation of the Data phase started in week-2.
- Overview about scoping a project.
Who should take this course?
This course is directed towards people comfortable working with Jupyter Notebooks. Specifically towards the people who want to use the model they developed in production. It discusses the need to step out and look at the overall processes involved in a Machine learning project in production environments. Since this is a course aimed towards freshers, most of the course content would probably be repetitive if you are a working professional.
How do I rate this course?
This course is that it is entirely theoretical. As this is a part of the specialisation, we are first introduced to the theory, and the rest of the courses focus entirely on the practical aspect. However, A 3-week course being entirely theoretical makes this course feel dry.
One other thing to note is that, in the follow-up courses, week-1 will discuss the same things discussed in this course. They start again with challenges in specific lifecycle phases and then expand upon them.
What else is there to know about this course?
Well, the selling point of this course is Andrew Ng. He explains the concepts through different examples, making everything easy to understand. However, it is understood that this course is only a way to prepare people for the following specialization courses.
How has this helped me?
Well, for starters, I took this course right around my job interview for the position of ML engineer. I speed-ran through the course and made brief notes summarising everything. This has helped me understand what I needed to do as an ML engineer and how I needed to answer a few questions asked in the interview. (P.S. I got the job.)
So, in that sense of getting an overall idea of the field, this is a wonderful course!
If you have a decent modelling background, then you should take the courses from course-2 of this specialisation. Those courses deal with different practical tools that can be used in production environments while also discussing everything from Course-1.
If you want the specialization certificate, then you can take this course at the end of the specialization and complete the quizzes.
That's it for the review of this course. If you still need to decide about taking this course, I recommend you gloss over my articles based on this course here.
If you found this review helpful and are interested in learning more about AI and Data Science, check out my other articles on NeuroNuts. Sign up to join me in the AI & Data Science journey! You can also follow me on Medium.