Hi, I'm a Data Scientist / ML Engineer and many people ask me the same question: How do I get started on learning AI / Machine Learning? I'll try to give you a quick answer here as I see that this topic is also hot in the Indie Hacker scene.
First thing to know: you'll first learn Machine Learning and then Deep Learning in case you want to go further. AI is only a marketing word right now. You'll discover it once you get into ML and DL.
So, what should be your starting point to learn ML?
Summarizing, if your background is technical, you should be more than comfortable to start on learning ML.
Which are the best resources to learn ML?
How could you transition to DL?
My book in this case would be "Deep Learning with Python" by Francois Chollet. He's the creator of Keras, one of the two most famous DL frameworks nowadays. You could also follow the rest of fast.ai courses, which use another framework called Pytorch (fast.ai is also a library built on top of Pytorch which makes things simpler).
In case you have any question, please ask me. You can also do it on Twitter DMs if you prefer (twitter.com/rodbarest). I hope this helps :)
Just in case people are also looking for courses, make sure to check out these platforms too -
Hi, thanks for adding them! I've always heard good feedback from at least courses 1 and 3, but as I haven't taken them I didn't want to recommend them.
Oh okay!
Do give it a go anyways! Never know what you might learn from it haha!
Hi @rodbarest, thanks for this post. From your perspective, how fast can an aspiring ML/DL engineer with strong software development skills actually start adding value to an AI product or create one himself?
Maybe it's my ignorance on this kind of work but building stuff like GPT-3 seems so daunting while on the other hand, one can just actually use the GPT-3 API with little to no background in data science (I think) to create a product.
Hi! Having strong SW development skills puts you on a very interesting position within the "ML ops" world. Data scientists normally don't have an advance knoledge of SW skills and when it comes to put models into production this is a problem. I think that if you start learning the basics of ML/DL and you start practising with some modelling tasks, you'd be ready to make a difference in ML ops tasks.
If you like specifically the ML/DL modelling part and want to understand what's really under the hood, you should maybe first spend 1 year or so studying it (assuming you also have another job).
The machine learning market expected to grow from $1 Billion in 2016 to USD 9 Billion by 2022, at a CAGR of 44% during the forecast period according to Market and Markets
Learn more about ML and its applications here: https://bit.ly/3tVSfA9
I don't like Maths. I don't find Statistics, Algebra, Calcus , Probability interesting. I find irritating to solve such problem statements.
With that said, is ML good enough to jump into for a person like me ?
Easy it better off just focus on Natural Language Processing to enter the world of AI ?
My Background
I guess in that case you'll have to focus more on building products based on existing ML/DL models rather than creating your own models. That means for example using GPT-3 to create a text generation product or using some Computer Vision model to extract objects from images for example. There are some models which are developed by the research community and thus you can be sure that they'll work fine.
So you could totally focus on the product side of things, and of course there are plenty of opportunities. However, it would be harder for you to get into a data science role on a company for example.