Hi, and welcome to sefisoft career Path on Machine learning and Artificial Intelligence Today, I’ll be taking you through the path to becoming a Machine learning and AI expert.
Now, broadly, there are two ways that I normally Describe it, if someone comes and asked me, I’m there a lot of questions that I would ask and you would need to rate yourself.
You’d need to have a very clear understanding about your capabilities to you know, understand, which is a way that you should approach.
So broadly, there are two Path I normally call
- A Math way
- The code way
So Math ways Path something that I did when I wanted to get started in deep learning on Machine learning.
1. Math Way.
So Math way, someone is kind of suited for someone who’s coming from Maths and statistics kind of background.
You don’t have as much good understanding about how coding is done, you don’t have as much and good understanding about software development,
So firstly, you have to understand the Maths behind all of this Machine learning Algorithms. And you have to get that right.
And you can use that using online free resources, you could take up courses, whatever way suits you, that’s perfectly fine.
But the first step to get in the Maths part is get your Maths correct and absolutely correct, right. Because that’s something that’s what we would be sort of leveraging for our rest of the path.
Once you get your Maths up part correct, then you have to start on kaggle challenges, obviously, you are going to get stuck up.
I mean, I remember the days when I wanted to get started on this problems, and I would probably spend a complete day trying to figure out how to implement even the most basic of linear regression or logistic regression, because it’s something I was not very familiar with.
And so that’s gonna be part of both the journeys, I mean, you have to start at a problem gets stuck.
If you’re not getting stuck, if you’re just looking at you know, guided implementation or any of that, probably your chances, it’s gonna take you a lot longer for you to actually get hands on in Machine learning.
So that’s I mean, that’s a sort of common advice that I give to everyone that you know, you have to start with a problem, get stuck, and try and solve it by yourself.
Don’t try and look at solutions. Try and, you know, try and do read up Documentation. So for example, if you’re coming, if you’re following the Maths way, then if you’re stuck in a kaggle problem.
Then the first step that I normally suggest is look up the Documentation. So if you’re using PyChart, or Kara’s or whatever the library you’re using, look up the Documentation. And from there, try and figure out what is a way to you can solve the problem, Once you have that part solved out.
And you can you can also basically look at kaggle has a lot of this open kernels that are present, right, you can look at some of those Codes, to understand how other people have coded.
in the Math very is important for you to understand why an Algorithm is working, and Why an Algorithm is not working.
And what is probably the ways you can resolve that. And you only look at Python open kernel implementation just to get an idea about how to write the code.
Don’t try To copy it, whatever the processing, and all of that they have done, your idea of looking at open source implementation is just to understand how to write the code.
Once you familiarize yourself with building hands on problems. And the next step is solve opens practical problems.
I mean, these are not kaggle as a as data sets as which are very, I would say clean data science data sources.
So don’t try and work on kaggle problems, try and work on more real practical problems, you can get them on other websites, you can do some freelancing work, do whatever you want to but get some hands on real data, right?
kaggle data is sort of very fabricated clean data most of the times. So once you do work on practical problems, then you can start blogging about them, And then probably publicizing your work.
2. Code way.
You don’t need to understand the Algorithms on the day one, that’s perfectly fine. If you just get yourself familiarized with what a psyche does, what the Kara’s does and how they solve a problem, right?
For example, if given data set, how you run a random forest Algorithm or how you run a decision tree Algorithm, just get familiarized with that.
This is again important in this part of the journey, as well that you have to start on a problem and get stuck, you would have Algorithms you would probably because you’re someone who’s coming from software part of things.
you probably would not get stuck in quoting up the entire thing, but he would get stuck in having very low accuracies right because, he would not know what is what is an Algorithm working out ,how is an Algorithm working on probably,
what are the steps that are missing up so To do that, you then look at what are the other people who have solved the same problem on kaggle? How they have done it right, and understand why they have done it.
So that’s how you approach this particular pathway, which is that you don’t get stuck and understand why are people getting stuck, where they’re getting stuck?
And how are people solving other people solving the same problem and understand the merits or demerits?
And that’s how you start getting familiarized with the Math part of things which you have to do in any ways, right?
So after you do that, then again, as I said, you start with practical problems once you are done with the kaggle problems, because we need to solve real life problems.
we can just solve clean data sight problems, like kaggle, why does it solve practical problems, then you can blog about them publicize your work.
And that’s, that’s broadly how you can complete this journey. So broadly, both of these journeys have slightly different initial journey that is very different.
But after that, you have to start working on practical problems, make a blog about them publicized about them.
If you know some path to become machine learning expert then you can comment us below in comment box.
