Honestly, for your situation, I would suggest thinking about it like this: Brilliant.org Premium Subscription is for building the mental scaffolding, while Coursera Deep Learning Specialization is for actually laying the bricks.

I've spent a lot of time on both and here is the breakdown:

* **Brilliant**: It's highkey better for fixing that rusty math. It won't teach you to build a production-ready model, but it makes the 'why' behind backpropagation feel like a game. It's great for building confidence so you don't quit when things get heavy.
* **Coursera (Andrew Ng)**: This is the gold standard. If you want to actually code a neural network in Python from scratch, this is it. But yeah, it's dense and can be a slog if you aren't in the right headspace.

My advice? Start with a month of Brilliant to wake up your brain, then jump into the Coursera Machine Learning Specialization once you feel 'math-ready.' It's a safer bet to prevent burnout. Good luck!

Oh man, i feel u on this! I was literally in the exact same spot a few months ago. I started with Brilliant.org Premium Subscription because I love the gamified vibe, and honestly, it's AMAZING for fixing that rusty math. It basically made linear algebra click for me in a way college never did! But i found that after a while, I kinda wanted to actually build stuff, you know?

So yeah, here is what I recommend:

- Start with Brilliant.org for like a month to get ur confidence up with the logic and math puzzles. It's sooo satisfying!!
- Then, switch over to the Coursera Machine Learning Specialization by Andrew Ng.

Basically, Brilliant.org teaches you how to think, but Coursera teaches you how to actually code the models in Python. Ngl, the academic stuff on Coursera can be a bit dry, but having that certificate is a highkey mood booster for the resume. If ur budget is $20-$50, you can definitely swing one at a time. Good luck with the journey!!

tbh the biggest trap is staying in 'tutorial hell' for too long. Whether you pick the gamified route or the academic one, youre working in a sandbox. Industry standards in AI development require you to actually manage your own environments, handle CUDA versioning, and deal with messy data that hasn't been cleaned for a lesson. TL;DR: Use these for 30% of your time; spend the other 70% failing to build something locally. My warning is this: dont confuse finishing a curriculum with actual engineering competency. You can know the theory behind gradient descent but still have no clue how to debug a shape mismatch in your tensors. If you want to transition into a real role, you need to get comfortable with the raw documentation and GitHub. These platforms make you feel smart, but they shield you from the 'dependency hell' and infrastructure hurdles that define the actual job. Start your own repo on day one and try to replicate what youre learning outside of their browser-based IDEs.

Lol I was literally about to post the same thing. Glad someone else brought it up.

Any updates on this?

Ok so, I’ve been around the block with different platforms for years, and even though I’m just starting with AI specifically, I’ve noticed a huge split in the market. It's basically a battle between the "University" model and the "Gamified" model. From what I've seen, here is how they usually play out: * The academic sites are selling prestige. They want you to feel like you’re in a real classroom, which is cool, but *so* easy to fall behind on if you have a busy life. I tried the heavy video route once and basically just had them playing in the background while I scrolled my phone.
* The interactive brands are selling "flow." They’re designed to keep you clicking, which is great for staying awake after work, but sometimes I worry if I'm just getting good at *their* specific puzzles rather than actually learning to code from scratch.
* I actually found that using a mix of free "notebook" style websites (where you run code in the browser) and just watching long-form walk-throughs on YouTube was a better vibe for me. Honestly, have you checked out some of the free community-driven roadmaps? Tbh, I learned more by trying to break a simple project I found on a code-sharing site than I did from any paid subscription. It's less about the brand name and more about where you'll actually spend time typing code.

Yep been there done that. Can confirm everything said above is spot on.

Ok adding this to my list of things to try. Thanks for the tip!