How I Started My AI Journey as a Developer (Without Knowing What GPT Even Means)

As a full-stack developer, I spend hours every day writing code, fixing bugs, testing features, and managing tasks. Somewhere along the way, I realized I was repeating myself—writing similar functions, Googling the same snippets, and setting reminders for routine tasks.

But

At the same time, I kept hearing things like:

“AI will take developer jobs.”
“If you don’t learn AI now, you’ll be irrelevant tomorrow.”
“The future belongs to AI—developers who don’t adapt will be left behind.”

---

Honestly, it was overwhelming. I wasn’t a data scientist. I am just a simple developer.

I don’t know what is models, LLMs, GenAI. But I knew how to build things. So I decided to stop overthinking and just start exploring.

Where I Actually Began – With Curiosity, Not Theory

I didn’t know where to start. With a lot of enthusiasm and curiosity, I just started. As any developer, I go to Google, YouTube, and ChatGPT to help me learn AI.

But

It is not easy… there was so much information, tons of articles, but most of it was packed with heavy terms like transformer, tokenization, Ollama, Hugging Face, fine-tuning RAG, and so on.

My goal wasn’t to build the next ChatGPT; I just wanted to understand the basics. What exactly is GPT? What is OpenAI? Everyone kept saying "AI, AI, AI," but I needed to know the fundamentals first.

By luck, I found a course link from Chaicode.com by Hitesh Sir and Piyush Sir, which was focused on GenAI with Python. My good luck!

I had never purchased anything before to help with my career. This was my first investment, but I knew it was going to be worth it. Their guidance on YouTube had already helped me a lot, and now I felt confident that this course would help me learn AI.

That’s where my journey truly began.

---

Understanding the Basics

Once I started, I realized that to build anything useful, I first needed to understand some core concepts. It was shocking, but true—I finally got what tokenization, transformers, and self-attention actually mean.

I used to thinkg ChatGPT and the OpenAI API would give the same answer to simple questions, like “What’s the weather in Delhi?”

But I was wrong

ChatGPT gave me the answer. But the OpenAI API said... nothing useful. Why?

Here is the simple reason

🧠 ChatGPT is like a smart robot with internet and extra tools.
📦 OpenAI API is like the same smart robot, but without internet or tools.

In simple terms, ChatGPT is like a brain with a full working body — it can think and take action. It can search the web, do calculations, read files — just like a body helping the brain get things done.

But the OpenAI API is just the brain.
It’s super smart and can think really well, but it can’t do anything by itself.
If you want it to act — like search the web or get the weather — you have to build the body around it..

Wrapping Up

So that’s how my AI journey started with small steps, a curious mind, and a strong developer instinct.

From feeling overwhelmed by AI buzzwords to understanding how tools like ChatGPT actually work

And trust me, if I can do it, you definitely can too.

In the next post, I’ll break down one of the most basic but powerful concepts in AI:

What is tokenization?
(Hint: It’s like teaching a computer how to read!)