From 10 MB a day at sea to AI engineer

Life onboard

I finished a bachelor in marine engineering in Rijeka, Croatia in 2017 and went sailing as an engineering cadet right after college. I went through training and boarded my first ship in Tokyo. I was very nervous, but also excited because I knew this would be a huge challenge and I wanted to see if I could do it. Growing up in a port town, I’ve heard countless stories about life on board a ship from my dad and others. I knew it was hard, but that it was exciting and adventurous.

Myself and 11 other cadets boarded the 300 meter ship. We left our stuff in our cabins and went on a tour of the ship with our instructor. Upon arrival at the galley (ship’s kitchen) the instructor mentioned that we could fill our drinking water here. I raised my hand and asked “Is this the only place we can get drinking water at?” and he said “Only here, and in two more places on the whole ship”. At that moment I realised that there’s no way in hell that I’ll be doing this job for the rest of my life. It took less than 2 hours since boarding for me to decide that I needed an exit plan. And an exit plan I had.

In the engine room, MOL LNG Transport (Europe) Ltd.

Exit plan

During my thesis I used MATLAB for calculations and got hooked. A friend then pointed me to CS50, a course that quietly changed my career trajectory. Between finishing the thesis and boarding my first ship, I completed CS50, learned Python, and started building projects purely for the joy of problem-solving.

So when I was packing up my things to go on board, I knew that the internet on board would be scarce to say the least. We had 10 MB per day of internet traffic to be precise. Therefore, I downloaded as many books, courses, and videos as I could so I could do coding in my free time as a hobby on board.

In that moment in the ship’s galley, when I realised I needed an exit plan, I knew that learning to code wouldn’t be a hobby anymore, but my sole focus in life for the time being.

My stint on that container ship was 3 months and I spent almost every waking hour of my free time studying and working on small projects to learn as much as I can. When I came home in January 2018, I discussed with family and friends and realised the smarter thing to do would be to finish my cadetship, get my engineering officer certification, and then switch careers. And I’m glad I did it that way. It might sound corny, but my time at sea really toughened me up both physically and mentally. I experienced so many amazing things that so few people ever do. I met so many amazing people, travelled to so many places, and learned a lot about life!

Over the next 2 and a half years, I went on board 5 more times. In my 4th stint, I was determined to get serious about coding and switch careers ASAP. That tour was the worst one. Regular 16 hour work days, being woken up in the middle of the night because of engine failures on a regular basis, frequent watch-keeping, storms, and even COVID hit while I was on board. It was really a perfect storm of circumstances that would make anyone put coding on pause.

But I stuck through it and I pushed myself as hard as I could. Why? Because the alternative in my mind was having to do these types of stints for the rest of my life. And I would be ashamed of myself if at 70 years old I would look back and think to myself “I just needed to work harder and my life would’ve been much different”. In my mind this was life or death. So I grinded as hard as I could through that tour. Without exaggeration, on a couple of occasions I even fell asleep at my laptop and I would wake up and continue coding. When I came home I was exhausted, but I didn’t want to stop. I continued coding and studying at least 8 hours every single day, even though I was on vacation.

Then came my last tour. Before that stint I told myself that this was the last one. I knew I learned a lot so far and that I had around 3 more months on board to fill gaps in my knowledge.

When I came home, a friend recommended me for a junior developer role. I did the interview and got the offer. I was ecstatic. I couldn’t believe that all that hard work and suffering had finally paid off. I quit my job as a marine engineer and started working as a developer. Finally!

Moving up in the development world

As most people in tech, I watched the iconic series “Silicon Valley”. I watched it when I was sailing and I needed to unwind. I was dreaming of finding a job at a company like Pied Piper. Some people might cringe at this, but when you’re working in an engine room that’s constantly at 40°C to 55°C (104 F to 131 F) doing hard manual labour, the thought of working at a nice air conditioned open concept office is a dream. Not only that, but doing what you’re passionate about as well? Now that’s the dream!

However, my first role as a developer from 2020 to 2021 wasn’t quite like that. To give you some perspective, we didn’t use Git and we were emailing code files to each other. Not quite “Silicon Valley”, right? But nevertheless, I learned a ton and woke up excited every morning to go to work. Plus, I had the pleasure of having the best boss I’ve ever had (shoutout Jeremy!).

After a year at that company, I got a chance to interview for a newly opened company in my town of Rijeka. They sounded promising and exciting! They used React and Node, and they had a proper development team of around 40 people. I went through a couple of rounds of interviews and got the offer in September 2021. It was an open concept office, it was air conditioned, and I joined the backend team which had the most cracked developers I ever met. Now this was my “Silicon Valley”!

I cannot begin to describe how much I cherish my time with that team. I’ve made lifelong friends. I learned so much from every single person in that team that I’m indebted to them for the rest of my life. This was also the time I saw the most improvement in my skills as a developer. In those 3 years I worked constantly with Node, Next, and AWS and learned so much about frontend, backend, QAing, infrastructure, and everything I know about the software development process (which will come in handy later).

New challenge

In 2023, during my second year at my dream job, I was cruising. I was good at my job and I was loving it. I was loving life in general! Around that time I met my girlfriend who I’m still with. I did a roadtrip through the Balkans for about 2 months and worked remotely. I travelled the length of Japan for a month and a half, all the while working remotely, and I did a bunch of odd weekend trips here and there. During summer I had so much free time to enjoy the beach and to hang out with friends. Everything seemed perfect! I was living the dream.

But there was something missing. I needed a new challenge. On a few occasions during 2023 a friend mentioned that he joined a mastermind group that helps and guides people on how to start their own agency. He was getting amazing results and was pushing me to join because he thought that this was perfect for me. I had my reservations since I knew starting an agency would take a lot of work and I was pretty content with my laid-back lifestyle. But the idea of starting my own business was just too appealing. I took his advice and joined.

I opened my agency in September 2023 while I was still working at my day job. The plan was I would be doing 40 hours as usual in my day job, and at least 25 hours working on my agency. I thought it would be manageable and that I only needed to endure around 3 months of that schedule, and by that time I would have at least a couple of clients so that I could quit my day job and focus solely on my business. Boy, was I wrong!

Remember how I said that I had reservations because it would take a lot of work? It took a lot of work and then some! I was doing regular 70–80 hour weeks. The reason for that is that I needed to learn more skills to be a successful agency owner. I already knew development, but to be a successful CEO, you shouldn’t be the one doing the actual coding. So I needed to learn how to hire people. But to hire people I needed to get projects, and to get projects I needed to learn sales. But to even get people on sales calls, I need to learn lead generation.

I was lucky that I started my agency when I did because ChatGPT was released just a year earlier and everybody and their mom wanted to build an AI web app. Luckily I was there on Upwork and Fiverr to help them do exactly that. Slowly but surely I got better at lead gen, sales, and hiring, and closed and delivered projects for clients in the US, UK and EU, including document search for TCE that cut the time engineers spent searching for requirements by an estimated 83%.

Learning sales was the most difficult part. I was kind of a socially anxious guy who hated the idea of having to sell something. I watched so many training videos, courses, read dozens of books, practiced for hours on end, until I got to a point where selling became more natural. I don’t think I’m good at sales even to this day, but I managed to get to a point where I feel comfortable on sales calls. I managed to get to that by not following any set framework posed by sales gurus, but putting my own spin on it and just being honest and genuine. However, that’s a topic for a different blog post.

I can vividly remember closing our biggest project at the time, an AI document search chatbot for TCE, a New York infrastructure company. I couldn’t believe that it actually happened. I was blown away by the fact that just 4 years ago I was still sailing and dreamt of working as a developer. I had the same feeling of happiness and disbelief as I did when I got the offer for my first dev job.

At a certain point, I was drowning in too many projects. I couldn’t keep up with sales and delivering projects at a quality level I was aspiring to. It was a good problem to have, but I needed help. So I turned to a close friend and mentor who I worked with at my “Silicon Valley”-esque job. He joined as a partner and CTO in December 2024 and I can confidently say it was one of the best decisions I made in my career.

What three years of running a studio taught me

I first published this story on Medium in October 2025. A year later the ending is different, so I’ve rewritten this part.

Running a studio meant doing every job, from lead generation and sales to hiring and delivery. Doing all of them showed me which part I’m best at. It’s building AI that fixes real problems, close to the people who use it.

I stayed through launch on every build. Selling what I build keeps my promises realistic. Building what I sell means I know how to prove it.

What I want next is one product and one strong team. I want to go deep, helping many customers succeed with the same product instead of starting over each time.

So I’m looking for a full-time role. I’m open to fully remote AI engineering and AI solutions roles, and I can start now. My CV has the details, and my email is [email protected].

The engine room still shapes how I work. Understand the system you are operating in before you change it, and make sure the person responsible can verify the result.