According to the KOD publication from Yandex Practicum, 2023 became the year of neural networks. The trend of using technologies based on them continues this year and influences frontend development. Oleg Akbarov has been in this field for over 20 years and currently uses AI in his startup. He talked about his path in the profession, notable projects, and trends in the industry.
Oleg, how did you get into the profession? Why were you interested in development in particular?
As a child, I was a big gamer and wrote modifications for Warcraft 3, but a career as a programmer didn't really appeal to me. I loved the movie "The Matrix," in which the main character, the hacker Neo, hacked computer networks at night and disappeared at underground raves, and during the day was a simple programmer, a prisoner
of an office cubicle. That's how I imagined a programmer's career. The idea of being on a first-name basis with a computer always intrigued me, but the prospect of wearing a suit in the office and writing software for corporations didn't appeal to me very much (I'll note in the margins that fate is not without irony, and I still wrote my share of software for banks). Once, friends dared me to create an online store for them - this became my first project, and I must say, the process was captivating. After a couple of years of freelancing, I decided that it would be interesting to work in a team and got a job at a bank. At that moment, it became clear - development is here to stay.
Even as a child, did you realize the promising nature of this field? Of course, as a child, I didn't think about such things. I followed my curiosity and intuition and tried to find something that could be "mine." Moreover, I still get a little offended when I'm called a programmer. I know how to program, but a frontend developer is always a bit of a designer and UX specialist, and a good frontender is also a product manager.
If I talk about the moment when I realized that my career in IT might last, I think of 2012, when Instagram, Airbnb, and Uber were gaining momentum, and it seemed that almost all problems could be solved with an application. I was captured by the idea that one program, one service could affect the lives of millions. This thought was a beacon that helped me navigate early in my career: I didn't want a high-paying job, I wanted to do something that everyone knows and uses.
That's amazing! Tell us about the most notable projects you've worked on in the last 20 years.
I have many good memories of my time at Replika. An incredibly talented team and an equally ambitious idea. In addition, we were seriously ahead of our time, developing artificial intelligence that would benefit people. From a technical point of view, the web version of the service, which I was working on, used all the latest achievements in web development at that time: it was one of the first major projects that used the React/Redux/TypeScript stack. We worked closely with the developers of these frameworks and programming languages.
At Revoult, I gained experience working with huge codebases and a team of hundreds of engineers. I was responsible for launching the web application in the key US market, and my team and I were able to do the impossible, meeting extremely tight deadlines. Without a doubt, the launch in America affected the company's value, because a year later it rose to $40 billion. And, of course, the most interesting project is the one you're working on right now. Perplexity AI is a new generation search service that answers user questions, using all available information from the Internet. We have completely changed the process of searching for information using the latest advances in AI and large language models LLM.
What industry trends can you, as an expert, highlight?
Of course, AI is the main trend. In reality, it's more of a paradigm shift in how we build digital products than a short-term fad. Large language models LLM are capable of things that were unimaginable just a couple of years ago, and most importantly, every success in the scientific foundations of this field adds new tools to the hands of developers and researchers. What started as text generation continues in the form of images and videos, and in the future will create everything that can be represented as zeros and ones. That is, literally everything. It's important to understand that this is good news. Here's an example: if before, to make a movie, you needed a studio and actors, then very soon you will only need an idea, good taste, and a little computing power.
Oleg, can you share advice for those who are just planning to take their first steps in your field?
