Interview with Mirel Gugila – AI Senior Developer: “AI doesn’t replace humans – it transforms them into orchestrators of solutions.”
#ForTheLoveOfProgramming – From the AI Perspective
With over eight years of experience at Roweb, Mirel Gugila is part of the generation of developers who grew up alongside technology and witnessed the spectacular transformation brought by Artificial Intelligence.
Outside the office, Mirel is passionate about bodybuilding – a hobby he discovered in 2019 and one he uses as a way to relax and maintain mental balance.
“For me, the gym is the place where I reset myself. It helps me clear my mind and return to work with more energy and clarity,” he says with a smile.
In this interview, Mirel shares how his professional journey began, how he discovered his passion for AI, and how he envisions the future of programming in the age of automation.
How would you describe your journey as a developer — where did you start, and what attracted you to this field?
I started at Roweb more than eight years ago, while I was still in university, at 22. I met Florin Obogeanu, the coordinator of Roweb’s Craiova office, at an event hosted by the Faculty of Automation.
Although I initially attended another company, my conversation with Florin was so engaging and relaxed that he convinced me to join Roweb.
I began with a three-month internship, but by the second month, I received an offer to stay. Since then, Roweb has been my only workplace.
When did you feel like you started to “gain momentum” as a programmer?
The key moment came after my first year when I began working more independently.
In the beginning, I was part of a team with many senior developers – an experience that helped me grow very fast.
After a few months, I realized I could handle complex problems on my own and truly started to think like a “real programmer.”
One of Roweb’s biggest advantages is the work environment – you collaborate with experienced yet easygoing people who are always ready to help and encourage experimentation.
How long have you been with Roweb, and what roles have you been involved in?
I’ve been at Roweb for more than eight years. I started as an intern, then became a junior developer, and today I’m a senior developer involved not only in development but also in mentoring new interns who join our program.
What do you like most about Roweb’s work environment?
The atmosphere. It’s a perfect mix of professionalism and friendship.
We have great team chemistry, fun extra activities, and people who are both relaxed and dedicated.
My mentor was Ion Ghitun (John) – someone who truly inspired me during my early years here.
How did your interest in Artificial Intelligence begin?
About a year ago, a client from London challenged me to use AI in a project that involved processing a massive volume of CVs.
Previously, a team from India manually analyzed these CVs to assign internal classification codes – a slow and error-prone process.
We started researching and found ways to use AI to interpret CVs automatically and generate the correct codes.
In the end, we replaced a five-person team with a much faster and more accurate AI-driven system.
What did you discover through that AI experience?
I learned that prompt design is everything.
For example, at first, the AI was miscalculating people’s professional experience because it didn’t “know” the current date.
It was a funny but time-consuming problem. Once we realized the issue, we simply added the current date to the prompt – and the results became perfect.
That experience showed me how crucial it is to communicate properly with AI and how much logical structure is needed when designing prompts.
How do you integrate AI into Roweb projects today?
Beyond that initial CV analysis project, we started building, together with our client, an AI platform integrated directly into the development environment.
The goal was to reduce implementation time and deliver cleaner, faster code.
The results exceeded expectations – development time dropped significantly, and AI now allows us to focus more on architecture and product quality.
What is the most difficult part of programming with AI?
You need to be a very solid developer.
The challenge isn’t just coding; it’s about documentation.
AI needs precise specifications – if the documentation isn’t complete or well-structured, the result can easily diverge from what you intended.
So you must design the architecture carefully, structure your prompts clearly, and always verify the output.
How do you manage AI-related risks, like bias or lack of explainability?
Through testing and caution.
I test extensively, use detailed prompts, and never allow AI to take full control.
It’s important to stay one step ahead of it – to understand what it does and why. In the end, the human developer remains responsible for quality.
What are the main advantages AI brings to projects and businesses?
Speed, without a doubt.
We can now develop solutions faster, identify bugs almost instantly, and cut down work time by 5–6 times in some cases.
Before AI, I used to manually check hundreds of files; now, AI points out the exact source of an error instantly.
That leaves much more room for creativity and for focusing on what truly matters.
What are AI’s current limitations?
Prompts remain the key.
If you don’t phrase your requests correctly, AI can return entirely different answers to the same question.
Also, response times can be inconsistent, and that unpredictability can be frustrating.
So, even now, humans must remain in control of direction and validation.
Do you think programming will ever be fully automated?
No. AI will become an essential tool, but it will never replace human experience and critical thinking.
Programmers are – and will remain – orchestrators: people who see the big picture, understand what the final solution should look like, and adjust accordingly.
AI can execute tasks, but it still needs a conductor.
Do you think not having access to AI early in your career helped you?
Definitely.
It helped me develop strong logic, critical thinking, and a solid understanding of how systems work.
If you start directly with AI, you risk skipping the fundamentals and becoming overly dependent on it.
How have the skills required for today’s AI developers changed?
Today, you don’t necessarily need decades of experience to be effective – but you do need logical thinking, curiosity, and clear communication.
I use tools like Cloud Code and Cursor Editor daily, both AI-integrated, to better understand documentation and interpret client requirements faster.
Before, that process could take hours. Now, AI explains what needs to be done instantly – and even suggests solutions.
When it comes to AI-driven tools, I mainly work with Claude 4.5 Sonnet and GPT-5 High Fast – two complementary models that support different stages of development.
I use GPT-5 High Fast for planning, analysis, and structuring application logic, and Claude 4.5 Sonnet for actual code generation and refinement.
This combination gives me both speed and precision: I can quickly create a clear implementation strategy and then move on to writing clean, testable, and scalable code.
How do you see the future of programming with AI?
I believe that in the next five years, AI will be able to generate almost complete applications.
However, there will always be a need for human supervision and adaptation.
It’s like the Industrial Revolution – new tools will appear, but also new roles.
AI doesn’t eliminate jobs; it transforms them.
A message for young programmers or your Roweb colleagues?
Take advantage of everything technology offers – but don’t skip the learning stages.
It’s essential to understand programming logic, make mistakes, and grasp the “why” behind the code before relying on AI.
And of course – join Roweb! It’s a place with exciting projects, great people to learn from, and an environment that encourages continuous growth.
I was lucky to start here, and I wouldn’t change that for anything.
Outside of coding hours, I keep my balance through sports – I train regularly and believe the discipline you build in the gym directly reflects how you approach projects.
“When you learn to be patient and progress step by step, that’s exactly the mindset you apply in programming.”
