Review: LinkedInLearning course “Deploying Scalable Machine Learning for Data Science”

Review: LinkedInLearning course "Deploying Scalable Machine Learning for Data Science"

Today I am reviewing the LinkedIn Course entitled “Deploying Scalable Machine Learning for Data Science”.  The course is one hour and 43 minutes long and reminded me a lot of my time at university except with the added benefit of being able to pause the lecture and go and explore things of interest when they popped up. There is definitely a prerequisite in respects to knowledge. The tutor assumes that you are conversant with scripting, systems administration and associated concepts. Python and R are the basis for much of this course. I program in neither of these (being a C# programmer) but I was pretty much able to keep up.

The tutor is Dan Sullivan who has a sterling academic and working pedigree.

The Learning Objectives are:

  • Defining scalability
  • Tools and techniques for scalable machine learning
  • Architecture design patterns for scalable systems
  • Machine learning models as services
  • Containerizing models
  • Kubernetes for container orchestration
  • Monitoring performance
  • Best practices for scaling machine learning models

OK – I’m going to be blunt. At first I felt that this was all pretty dull. The first section of the course is entitled “The need to scale ML Models”. It has a lot of theory and not any “doing”. I’m a hands-on type of person so I struggled.

But then it started getting better. More code started going on the screen and I felt the need to pause the videos and pop over to Docker.com and download Docker and then go over and register with Docker hub. Yep, suddenly things started looking interesting because I could join in. I even took notes (I don’t normally take notes).

Conclusion

This is a good course but not for beginners – by which I mean beginners to IT. The ideal student, in my opinion, would already be savvy with Python and associated technologies because at some points you are going to want to jump into some scripting just to try out what the tutor is showing. The ideal student would also be familiar with systems administration because terms such as load balancing, autoscaling, performance monitoring etc are frequently used.

I am a programmer with a distant background in systems administration (Solaris, SCO Unix, NT 3.5 and 4 – yes I really am that old). My take on this course is that it is useful to both of these types of IT workers. So if you are either and you want to move into this area, then I recommend this course.

Thanks for reading,

Greg

0