My rough story of becoming an AI Engineer
Personal, raw, unfiltered story about my life
This will not be the usual Decoding AI article.
For once, I want to write something personal to better understand the story behind me as a content creator and AI Engineer.
As I was writing the text blocks for my new homepage, which will soon be released under pauliusztin.ai, I had to write the standard “About Me” section. Which is often the most challenging part, as you have to ask yourself, “Who really are you?”. I started with something generic, impersonal, but I wanted more. Thus, I ended up with a personal, raw, unfiltered story about my life.
I didn’t have the easiest path. Thus, as this is incredibly personal, it was hard to write.
I hope you will find it inspirational, get to know me better, or just have a laugh.
Here we go.
Since I trained my first neural network and built my first software app back in 2017, I have completely fallen in love with building software, especially AI software, that works.
I see my work as art, but instead of using a brush, I use my mouse, keyboard, and army of agents.
This passion for AI, software, teaching and sharing transformed into a clear mission to:
Build an AI Academy, but not as a huge business, but personal and optimized for you, where I want to teach people about building AI software that works.
Making knowledge accessible through AI, meaning I love building AI software that taps into the human collective and extracts exactly what you need, when you need it. That’s why I am obsessed with RAG, AI Agents and LLMs.
Here is how it all happened.
Born in 1996, in Timisoara, Romania, for anyone who knew me in the early years, nobody would have expected I would go the path I am on today. I was (and still am) an introverted person whose social battery is running out in hours. Also, I quite sucked at writing, speaking, or communicating. As a “fun” fact, I couldn’t read properly until I was 7 years old. But I was immediately attracted by tech. I was lucky enough that my parents bought a personal computer back in ~2000. Those big gray boxes running Windows 97. I didn’t know how to read correctly, but I started using the computer when I was 3 years old.
My life started pretty well. I lived in a good family. My father was an entrepreneur who created a successful chain of jewelry shops, while my mother was an artist. The dynamics in the house were positive, calm and full of hope. My parents loved me and offered me all the education I needed. I was always inclined toward the abstract world. Thus, my mother pushed me into maths and science, ultimately becoming a math olympiad.
Everything went amazingly well until 2008. During the financial crisis, everything changed. My father lost his business. Since then, it all spiraled down. We lost our house and hopes. I was in my teenage years at the time. This affected me hard. I started drinking every week. Then I discovered weed and ultimately harder stuff.
By some miracle that I can’t articulate, at 18 years old, I realized that if I continued down this path, my life would be miserable. I was living in Romania, lonely, with no money, no inheritance, no relationships and shitty energy.
At the time, I had my pre-university exams that I knew would set the path for the rest of my life. Thus, I rose, took the dust off, and did a complete U-turn in my life. I dropped my whole toxic “friends” group and started studying hard. I mean, really hard. 10-12 hours per day, every day. Ultimately, I aced my exams. Since then, I continued this work ethic through university, catching up on all the lost years on math, programming and physics. I was back in the game, at full focus.
In the 1st year of university, I participated in a hackathon, where I tried to build a mobile application. “Tried” is the keyword, because I didn’t get too far, but it lit a passion in me for building software.
My teaching career started back in 2016, tutoring my peers in advanced math and programming during my CS university years. Then, after I graduated, I continued this passion by teaching the foundations of AI laboratory at the Politehnica University of Timisoara.
During the 2nd year of university, in 2017, I started working for a local company as a software developer. Luckily, I had a fantastic team and mentor. I learned a ton about building production-ready software in Python and JavaScript. But the thing is, I knew I wanted to do AI. Thus, I had to move on. Through a university teacher that I learned Deep Learning from and got along really well with, I got a job at Continental, as an AI Researcher on autonomous driving problems such as 3D object detection and tracking. But that wasn’t for me. I fell in love with building, not researching.
As I started my journey from level -1, I had a difficult time working at companies that were successfully shipping AI to production. As I slowly went international and worked for more “mature” companies, I realized that finding a team that knows how to ship AI software is rare. As I wanted to learn how to build AI software that works, no matter the cost, I realized that from my position, the only solution is to start sharing online about my projects, findings, wins and failures.
That’s how I started my Medium and LinkedIn brand, then my newsletter, which ultimately morphed into the Decoding AI Magazine that we know today.
Meanwhile, as I knew I wanted to be more than just an influencer, but an actual AI engineer who helps the community grow, I started writing the LLM Engineer’s Handbook together with Maxime Labonne. I had 0 expectations when I started writing it, but I knew that the only thing I could do was to give my best. To my utter surprise, the book was sold worldwide, selling in one year 20000 copies, becoming a No. 1 Amazon bestseller and hitting the top 3 in Packt’s top sellers.
I was deeply grateful that people enjoyed my style of teaching, with an emphasis on understanding the whole picture: end-to-end AI systems, from data collection to shipping. That’s the only way you can learn to build AI software that works. You need that holistic view and understanding. And I was deeply grateful that the idea resonated with the world.
This achievement motivated me even further and made me start my own AI Agents course, together with Towards AI, teaching how to think, build, and ship AI Agents to production. With the same end-to-end, holistic systems mindset I adopted while writing the book. Still working on it. It will be done in late November. If you want, you can join the waitlist here.
With some luck and tons of hard work, I became a founding AI Engineer for a San Francisco start-up that solves the problem I love the most: Making knowledge accessible through AI.
Because this became my life, after one year, at the end of 2025 (which is now, as I am writing this), I started working on the 2nd edition of my LLM Engineer’s Handbook to update it with all my latest findings on building AI products. It will probably be done in ~April 2026.
This might look like a dream. To some extent, it is, as I still see myself as the introvert guy from Timisoara, Romania, who is too shy to negotiate his salary. But at the same time, I know I deserve it, because I put in the hard work, passion and determination required to get it. I didn’t accept failure as a result, but just as a chance to learn and move on.
Because I am not entirely a workaholic, mostly because of my beloved fiancée, Madalina, I love spending time with her and my cats. That’s what keeps me balanced and gives me the courage to do all of this.
This wasn’t easy for me to write (or share!). But I hope there are other people out there reading this who had a thought start, and this will motivate them to put in the hard work to get where they want to be.
See you next Tuesday with more posts on AI engineering!
What’s your take on today’s topic? Do you agree, disagree, or is there something I missed?
If you enjoyed this article, the ultimate compliment is to share our work.
How can I help
Join the waitlist for our latest course on Agentic AI Engineering, where you will learn to design, build, evaluate, and deploy sophisticated, production-grade AI agents and workflows. Done with Decoding AI’s builder mentality, the course consists of 30+ lessons, with code and theory, during which you will build a real-world AI product that you can show off during your interviews. Done in partnership with Towards AI.
The course will be released in late November.
Images
If not otherwise stated, all images are created by the author.
Thank you for sharing such a personal story, Paul. Your journey is inspiring and it does motivate me to keep going after what I am passionate to work in. Wish you the best :)
Reading this at the same age and with a similar life history motivates me a lot to keep sharing content and to do my best every day.