Why do so many self-taught coders fail or get stuck, lose momentum, or feel like they are not making progress even after months of tutorials? I have been there. Whether you’re learning software engineering, data science, AI, ML, WebDAV, or product engineering, this applies. I became a self-taught data analyst and then a data scientist, a whole long story, and I made plenty of mistakes along the way.
Today, I’m sharing four mistakes that I have personally made in my self-taught journey, and I wish I knew earlier. And then I’m going to share simple frameworks that you can use to move forward. Now, before we dive in, there is more than one way to learn.
The three ways that you have got is you go to school, the traditional route, you get a degree, study computer science or data science degree, and whatnot. Boot camps, which is short-term intensive programs, self-taught, building your own path with online resources, courses, tutorials, books, and projects. Each path has its pros and cons.
Degrees can offer more structure and credentials. Boot camps can speed up and make you focus. And self-taught gives you flexibility and ownership on your own learning, and it also saves you money.
But one thing we need to keep in mind that self-taught, becoming anything self-taught requires discipline. You have to be truly motivated from within to start on your journey and finish that journey. When you’re self-taught, there is no one chasing you.
You have to hold yourself accountable, which is why a lot of people get stuck in the current job market. Having a degree can sometimes give you a boost depending on the company and role. But that’s a topic for another blog.
This article is focused on how to succeed as a self-taught coder because it is absolutely possible. I am one of the examples. I’m a self-taught data scientist, and I wanted to share what worked and what didn’t work in my journey.
Here’s the first big mistake that people make. Information overload. The coding world is packed with content, especially today when so much information is available.
Online courses, YouTube video, blogs, newsletter, you name it. It’s really easy to get stuck thinking, where do I even start? What’s more important? What am I missing? And honestly, this was me when I started. I actually learned a lot of things on the job, which sounds good.
But the problem was I was only learning whatever the project required at that time. I didn’t have a big picture of what a full data science skillset looks like. So there were huge gaps I didn’t even realize I had until much later.
One thing that really helped me later on was learning about something called T-shaped learning. The idea is your skills should look like a T. The top of the T is broad knowledge across your field, understanding the overall landscape, the tool, the concepts. The vertical part of the T is deep expertise in one or two key areas.
For example, in data science, you might have broad knowledge of Python, SQL, statistics, machine learning, data engineering concepts, but then go deep into NLP or deep learning or experimentation, like pick one or two fields. When you are self-taught, it’s really easy to accidentally end up with an upside down T, super deep in one project area, but with big gaps in your breath. That was exactly me.
And honestly, this can happen no matter what path you take, even in degree programs, it’s possible to go deep in certain courses but miss the bigger picture. That’s why I think the T-shaped model is helpful for anyone. It’s not just for self-taught folks.
So a couple of tips to figure out your T-shaped learning style. First, narrow down your focus. Coding is too broad.
Are you aiming for software engineering, data engineering, machine learning, full stack? Once you have done that, then study the job descriptions that you want. Look on LinkedIn, what skills do people in those roles actually have? Those people whose job you actually want. That will help you build a proper T-shaped skill set instead of learning a random or only based on what your current project needs.
Number two mistake is the Classical Tutorial Trap. And yes, I have been there. So what exactly is tutorial trap? Let me explain.
You start watching tutorials one after another, Python, SQL, data structures, ML models, you name it. And after a while, you start thinking, wow, I have learned so much, right? But when it comes time to actually build something, your brain freezes. I’ll tell you why.
I remember this happened to me when I started I had watched so many tutorials. But when I had to code on my own, I almost forgot how to write hello world in Python. Okay, that’s exaggeration.
But the point is, when you actually start doing hands on work, you’re going to run into situations that were not covered in tutorials like coding is all about problem solving, you will solve one thing, you will end up in the next problem. So unless until you do it yourself, a tutorial watching a tutorial, watching a tutorial 2000 years later, and thinking that you have figured it out doesn’t work in reality. In our structure program, you would have projects.
But when you are self taught, no one is forcing you to practice. So it’s easy to get stuck in watching mode. One simple framework that has helped me break this cycle is learn to teach.
First, learn the concept, know when to stop watching the video, then actually do it build something, write code practice. And if you can explain it to someone else, even a blog post or just explaining it to a friend, it really locks in, do it. If you’re just watching tutorials without doing any practice, trust me, it’s not gonna stick, you actually have to do it.
Now let’s move on to mistake number three. The next mistake is a personal one.
It’s Skipping the fundamentals because you want to jump to the cool stuff, which is exactly what I did. Do not recommend. For example, you’re trying to build an AI model without understanding the basic math and stats, writing SQL queries, but not understanding join or indexes.
And I’ll be honest, this is one of the mistakes that I made. And looking back, it sounds so silly. There is something called a feature engineering in machine learning.
And it took me a long time to figure out what exactly a feature engineering is like I was doing it. I just couldn’t tie the name, the name of the concept to the actual work that I’m doing. Feature engineering is basically picking the right set of columns in your data that are going to be part of the model.
I was doing that. I just didn’t know that it was called feature engineering. And these are all the things that are covered in the fundamentals that I totally skipped over.
And that to date, I would say like, I still struggle with that. I don’t know if any of you relate to it, like, let me know in comments, or they’re like specific fundamentals that you know, maybe, but you just can’t put name next to it. These are some of the things that I feel like a lot of self taught coders, programmers kind of struggle with.
And one mindset that helps me in this is learn why not just how it’s easy to learn how to run a library or copy some code. But if you don’t know why it works, you will hit a wall when things break or when you need to customize it.
Last mistake that I am also guilty of is Not prepping for interviews.
People often forget that interviewing is a skill in its own. You might be great at building things. But can you solve coding challenges under time pressure, like in 10 minutes, explain your code clearly on a whiteboard to a stranger, walk through system designs or ML design questions.
It’s a lot. I used to think if I can do the job, I’ll be fine. It’s actually not that case.
interviews are completely different game and you have to prep for them. If you have goal to land that offer. And that is true for everyone, whether you are degree or you are self taught interviews is a skill of its own.
Everyone has to prep it. One framework that I use for this is content practice similar. So content is learn the material algorithm SQL ML system design.
Second step is you practice solve coding problems, mock questions, and number three similar do full mock interviews with a friend or a platform where you can practice that cycle really helped me get more comfortable for my first data science interview. And even today, these are some of the many mistakes that I have made in my self taught coding data science journey. And I want to know like, are there more that you would add from your personal experience or you wish somebody had told you? Let me know in comments.
And I hope you’re having a great day. I’ll see you in the next one.
Bye.