First Steps Towards Learning Machine Learning

Hey there 👋🏽,

A lot of my classmates/people who follow me on Twitter have been asking me how to get started in machine learning, so if you’re wondering the same, this guide is for you!

There are hundreds of guides online, so why should I read this one?

This is not meant to be an universal guide. It is for people who ask me how to get started or who have a learning style similar to mine. Regardless, it will take less than five minutes to quickly scan this guide, so why not do it and then determine whether or not it is helpful to you?

Nevertheless, I have personally found many of the online guides to be helpful. One I’ll mention is @cneuralnetwork’s Introduction to Machine Learning.

If you have a lot of experience with the machine learning environment, I would appreciate it if you could examine this page, point out errors or poor resources, or even recommend some new ones. It’ll help me keep the guide updated.

But before we begin, here’s something you should know —

Roadmaps suck

Yes, you read that right. It’s 2025, and the world is evolving rapidly. A new research is being published every day, especially in the ML ecosystem. If you are currently searching for roadmaps, I can assure you that even if you follow the roadmap precisely, you will be heading toward mediocrity.

Everything in tech, including machine learning, revolves around FAFO! Fuck around and find out! The best course of action is to remain curious and avoid doing things exactly as described, even while you are following this guide. This is not a Zero to Hero guide; rather, it is intended to serve as a foundation for someone just beginning their ML journey. Here’s a little secret: if you’re looking for a Zero to Hero guide, you’re not gonna make it!

If you want to read more about it, here’s a blog I recommend: Roadmapping Your Way to Mediocrity.

Prerequisites

A few things need to be checked before we start machine learning —

Python (obviously)

I believe most of the people reading this will already be somewhat familiar with the Python programming language.

Corey Schafer’s Python Tutorial: Here’s a 7-year-old Python tutorial by Corey, most of which still holds correct for today.
Kaggle’s Intro to Python Programming: Get started with programming if you have no coding experience.
Kaggle’s Python Course: A little more about Python.

The second and third resources, which are both from Kaggle and involve hands-on activities, are definitely worth looking at. If you need to learn more about a particular subject, check out Corey's Python tutorial.

Moreover, keep making small projects & games in Python along the way to reinforce your learning.

NumPy

NumPy is an open-source library that enables fast and efficient numerical computation in Python. Almost all the machine learning engineers & data scientists use NumPy, or something very similar to it, in their day-to-day lives. You’ll need NumPy from day 0 in your machine learning journey.

NumPy Documentation: The official NumPy documentation. Use it whenever you want to know about any class, function, or method in NumPy.
Getting started with NumPy & NumPy Arrays: It’s a blog I wrote some time ago, and you might find this useful.

In my honest opinion, you can just search for a NumPy tutorial on YouTube or the internet, which you can complete in a day. After that, try writing small scripts like matrix multiplication using NumPy.

Pandas

Pandas is a Python library that simplifies data manipulation, cleaning, and analysis for handling tabular data; it is widely used in fields like data science, machine learning, finance, and research. Learning it is a no-brainer.

Panda’s by Kaggle Learn: Kaggle’s hands-on mini-course is the best place to get started with Pandas.

Indeed, you do not actually need to know much about NumPy or Pandas to begin machine learning. Having foundational knowledge is fine, and then you can learn on the go. Above all, avoid attempting to finish those incredibly lengthy crash courses on these libraries.

Mathematics

Now here comes the most interesting part. Today's fancy tweets and groundbreaking research articles are filled with complicated mathematical calculations that are difficult to understand. To be honest, you’re not supposed to either, not at this point in your journey. I would strongly advise you to start with machine learning and then learn on the go rather than studying a lot of mathematics.

However, @cneuralnetwork’s Introduction to Machine Learning blog mentions a few really good resources for mathematics, so do check them out.

Again, if you have basic knowledge of linear algebra and calculus, that’ll be enough to start with.

Machine Learning

So it’s finally time to start with ML. To start with, here are two mini-courses that I would highly recommend you going through before starting any long course:

These two mini-courses by Kaggle are best to gauge your interest in machine learning.

Well, now that you know that you actually want to dive deep into machine learning, here are some courses you can do:

CS229 Machine Learning: This course is considered the goat. It covers almost all the topics one needs for native machine learning. Moreover, it’s led by Andrew Ng, one of the most renowned professors in the machine learning ecosystem. If you’re aiming to complete this course, then I’ll recommend you focus solely on this, as it has some pretty hard assignments too. All in all, if you’re ambitious, this is for you.
- Course lectures
- Notes, problem sets & solutions
Hands-On Machine Learning: If you’re an avid reader and prefer to learn through books, then you need to get your hands on this amazing book. The author not only explains most of the native machine learning topics but also dives deep into advanced topics like computer vision and natural language processing.
- Amazon link
Coursera Machine Learning Specialization: This is an ML specialization on Coursera, led by Andrew Ng. If you want to do CS229 but don’t really want to go through the effort of doing those lectures from YouTube and solving & grading those assignments by yourself, maybe this is the way. Although this comes with a price, you can apply for financial aid and complete this course for (almost) free. One more thing to note is it has simpler assignments and less mathematics compared to CS229.

‘Almost free’ because the financial aid depends from person to person. Some might get the course for free while others might have to pay a minimal amount of around ₹500.
- Coursera link

I will stop this list here so as not to overwhelm you with resources. Sticking to one of the courses listed here is sufficient; you are not required to finish them all. Also, this is not an exhaustive list; you’ll find many more resources on different blogs; these are some I have heard good reviews about. Additionally, since this guide is meant to be your first steps rather than a comprehensive roadmap, I have not included any deep learning courses or resources.

Conclusion

Before I end, please understand that you do not need to follow the resources I have suggested. Fuck around and find out what works for you and what doesn’t. Stay curious.

If you would like to read more from me, you can subscribe to The Neuron, a newsletter where I talk about machine learning in general, what I have been doing and also share some cool resources I find over the period. You can also follow me on Twitter/X.

That’s it for now. I hope you get started on your ML journey as soon as possible.

Thank you!