It’s been around a month and I promised to leave my thoughts on Educative.io after trying out the platform, so here goes…

I think Educative.io is a mixed bag. Since I joined a few months ago, I managed to complete “Cracking the Google Associate Cloud Engineer” and “Unleash the Power of Large Language Models Using LangChain” courses. I’m also working on “Grokking Coding Interview Patterns in Python” and “Optimization for Machine Learning with Numpy and Scipy”. Let me tell you about all 4 of these courses.

Cracking the Google Associate Cloud Engineer

This course sucked so much. Maybe it’s just that I hate these certifications from Amazon, Google and Microsoft. My last experience with cloud certifications was at the Federal Reserve Bank of Kansas City, where all of IT was encouraged to get at least 1 certification in a cloud platform. We all had free subscriptions to ACloudGuru, and… it’s just really hard not to scream F*CK out loud when I remember this. It’s like the saying you can lead a horse to water but you can’t make them drink it. For me, I hate the taste of cloud certifications and I just cannot do this to save my life.

Actually, I take that back. I did complete this course, but I have no plans on taking the test. Why? Because the videos sucked, the online quizzes sucked (you can just memorize the answers after failing a few times), the online labs are not up to date, you have to pay for said labs if you happened to use your free newbie credits years ago, oh, and did I mention that I hate these certification exams. Yeah, don’t take this course unless you’re a masochist.

Unleash the Power of Large Language Models Using LangChain

Ok, so this one wasn’t as bad as the last one, but it’s not really something I’d pay a big subscription for. Like other courses here, there’s a window where you can paste your code and run it, which is good, except you have to pay extra money to use their debugger (I refuse). There’s also a magic AI wand button next to the window, where you can hi-light text and get an explanation from a LLM. Unfortunately, the AI powered explanations are useless. If I didn’t understand something, I had a much easier time copying and pasting the text/code and talking to ChatGPT. Finally, nearly every part of the course is covered in greater detail in LangChain’s free documentation.

Now, I did get something out of this course. The code examples were good and it did motivate me to try a separate project (or was it my blog post that motivated me?). Overall though, I don’t think the course was worth the subscription. Moving on…

Grokking Coding Interview Patterns in Python

Now this course may actually be worth the subscription. Like the previous course, each lesson had a window and a useless magic AI button. Unlike the previous course though, this one is packed with content I did not find anywhere else. Prior to starting this course I had taken the “Data Structures and Algorithms Specialization” on Coursera by UC San Diego, and while the specialization did provide me the basics and some of the theory, this course gets you ready to do LeetCode style problems.

The one thing that’s missing from this course though are proofs of correctness. Yes, there are intuitive explanations for the solution, but I prefer to have a formal proof to go along with the intuition. That being said, I think this course is worth the subscription.

Optimization for Machine Learning with Numpy and Scipy

This is the course I’m working on right now. Like the last 3 courses, there’s the coding window and magic AI button. Unlike the last 3 courses, this course assumes that you know Linear Algebra and Multivariate Calculus, which I do. There was a section on Vector Calculus where I hadn’t seen this stuff since grad school, so this is not a course for beginners.

With this in mind, it’s not a math course either. There are only intuitive explanations and not many proofs in the course, which is pretty frustrating. I mean, if you assume that your audience knows Multivariate Calculus, is it really that much of a stretch to assume they also know how to read proofs?

Now, there is some good stuff I hadn’t seen before (or at least not in detail), like Newton’s method and BFGS, however, a lot of the stuff you can find freely available on the web. Like, it’s not hard to find an explanation for gradient descent and all its variations like RMSProp or Adam. So I’m gonna say this one is also not really worth the cost. I’ll be looking for some books or recorded lectures on optimization elsewhere.

So overall, there was only 1 course that was worth it. Besides that, Educative.io seems to be the only learning platform that has courses that prep you for the System Design Interview. Thus, if you’re not looking for help with the Coding or Systems Design Interviews, then I suggest you look elsewhere.