AI is everywhere. Companies are trying to integrate it into their products, and we are using it more and more in our everyday lives. But does that mean I should become an AI engineer?
For me, the answer is yes — but the reasons are more personal than simply “AI is the future.”
It started with Android
Let me take you back to 2013, when I received my first Android smartphone as a birthday gift.
At the time, I was learning Japanese and couldn't find an English–Japanese dictionary that was good enough for me. So I decided to create my own.
I learned Android development from YouTube videos and started building the dictionary application. I had countless questions about how to implement different parts of the app. Then I discovered Stack Overflow, and I felt like a miner who had just found a gold mine.
I was asking questions every day. Sometimes I asked such stupid questions that I even got temporarily banned. But I kept learning.
A few months later, I finished the dictionary application. It eventually became the subject of my bachelor's degree thesis.
After graduation, I started working as an Android developer. One of the first major projects I worked on was a government application that allowed people to interact with government services. It was called MyGov and is now used by millions of people in Uzbekistan.
Over the years, I worked on several mobile banking applications, including Milliy, Ipak Yuli Mobile, Joyda, and Davr Mobile. These applications are also used by millions of people and provide real value: people can perform many banking operations without having to visit a bank.
I was hired because of my experience, and I learned Android mostly by doing. But there was a problem: I never built a strong theoretical foundation.
For years, practical experience was enough. But when I started receiving interview invitations that focused heavily on theory, I struggled.
I don't have a problem with learning. My problem is that I forget information easily, and preparing for theoretical interviews feels overwhelming. Because of that, I had to reject several interview invitations.
It is difficult to admit that after so many years of experience, I still don't know the theory as well as I would like.
Then came ChatGPT
In 2022, ChatGPT was introduced.
I was one of the early users and started using it almost every day, even though it wasn't officially available in Uzbekistan at the time. I even helped several people create OpenAI accounts so they could use it.
At first, I mostly used ChatGPT for learning and coding. It was still quite primitive compared with today's AI tools. I would ask things like, “How do I implement this?” and use the answers to build things.
For example, I created an application to help me prepare for technical interviews. It worked both offline and online.
When I created the online version, however, I made a serious mistake. I wasn't experienced enough with security and accidentally left the database publicly accessible with default credentials.
An automated scanner eventually found it. The attackers encrypted the database and demanded Bitcoin to decrypt it.
Fortunately, the database didn't contain any private information. It only contained questions and answers for interview preparation.
Still, it was an important lesson for me: AI is a powerful tool, but I need to be responsible for what I allow it to do.
My experience with Cheapta
In 2024, I rebuilt the mobile application for Cheapta, my project for finding cheap flight tickets.
At the time, I decided to use Kotlin Multiplatform and Compose Multiplatform because I wanted to build Android, iOS, and web versions from the same codebase.
I successfully created the Android and iOS versions. The Android application was accepted by Google Play, but the iOS version was rejected by Apple.
The explanation was vague: the application wasn't using enough iOS system APIs, and Apple suggested that I turn it into a web application instead.
That decision frustrated me.
Why should I use an API just for the sake of using an API if I don't actually need it? I believe in simplicity: no unnecessary authorization, no unnecessary location permissions, and no unnecessary access to users' data.
You open Cheapta, and you simply get a list of countries you can fly to cheaply.
I sometimes wonder why developers are willing to pay $99 every year to build applications for Apple's platform, only to submit them for approval and wait to find out whether Apple will accept them.
Back in 2013, I paid Google Play's $25 developer fee once. Since then, I've submitted more than 10 applications, and almost all of them were accepted.
But the bigger problem came later.
For almost two years, I tried several times to create a web version of Cheapta using Kotlin Multiplatform. I kept running into problems. Eventually, I realized that KMP was creating more problems for this particular project than it was solving.
Then something interesting happened.
I created a web version of Cheapta in about half an hour.
That was the moment I stopped trying to force everything into one codebase. I realized that I could build a native application for each platform separately — and use AI to help me do it much faster.
From coding with AI to engineering with AI
Even a year ago, I mostly considered AI-assisted coding to be something fun to experiment with.
Then I started seeing professional developers use AI seriously in their everyday work.
I realized I should do the same.
Later, Google introduced Antigravity, providing access to models such as Claude Sonnet and Opus at a relatively low cost. I couldn't stop myself from creating projects.
I was amazed by how quickly I could turn an idea into something real.
And that's when I realized something important about myself:
I enjoy creating applications and solving problems much more than I enjoy writing code.
Ironically, I haven't written that much code manually over the last year.
AI changed the way I think about software development. The bottleneck is no longer always “How fast can I write this code?” It is increasingly “Can I clearly define the problem, design the solution, and use the right tools to build it?”
That's a very different skill set.
Why AI engineering?
AI isn't the future anymore.
AI is already here.
I can either adapt to this new reality or fall behind the industry.
But I don't want to become an “AI engineer” just because AI is popular. For me, AI engineering means much more than writing prompts in Antigravity or asking ChatGPT to generate code.
It means understanding how AI systems work, learning how to build applications around them, connecting them to external tools and data, evaluating their results, and ultimately using AI to solve real-world problems.
And strangely enough, this decision feels very familiar.
It reminds me of 2013, when I first started learning Android development.
Back then, I remember waiting for my university exams to finish because I couldn't wait to get back to my computer and continue building my own application.
Today, I have that same feeling again.
I want to learn, experiment, build things, and see what I can create.
The technology has changed.
The excitement hasn't.
That's why I decided to become an AI engineer.