From physics labs to AI breakthroughs: why the best AI architects think like scientists

“Experimental physics shaped how I think about AI”

— My interest in AI grew out of pure utility. While working at national facilities, I witnessed firsthand the transformative impact this emerging technology could have. But I also saw the real challenges of properly adopting AI in complex environments. Hands-on experience and the challenges of proper adoption guided me to my current role as an AI solution architect.

My current role as an AI Solution Architect at EPAM may seem worlds apart from my previous experience in experimental physics, but the parallels are striking. Working in experimental physics equipped me with a mindset ideal for pursuing rapid AI initiatives. It demanded deep software engineering expertise, collaboration within multidisciplinary teams, and the ability to deliver high-impact results — all of which are foundational to success in the AI field.

“Transformers unlocked a new world of possibilities”

— One of the most eye-opening moments in my career came when I started working with early deep learning models. At the time, solving even simple problems felt like climbing a mountain. Then transformers arrived — and everything changed.

Transformers weren’t just effective; they were “unreasonably effective.” Suddenly, we could tackle problems at a scale that was previously unimaginable. The first production-grade transformers, capable of running on standard machines, opened entirely new doors for research and innovation. They didn’t just improve existing methods; they opened entirely new doors for research and innovation. Driven by scale, we could only hypothesize what the next few years might hold.

 “Experiment, stay skeptical, and learn from trusted sources”

— When thinking about how AI will shape the future of work and life, my advice is to stay grounded and practical, especially in the short term. The tools and technologies we already have are incredibly powerful, yet we’ve only begun to explore their full potential.

Thriving in an AI-powered workplace is about staying curious and willing to learn. Join communities where AI knowledge is actively shared. Focus on areas where you feel least confident. Read from trusted sources, but always approach information with a critical eye. Stay skeptical, but don’t let that hold you back from experimenting. The best way to learn AI is through hands-on experience — it’s where the real breakthroughs happen.

I always recommend resources that focus on systems and principles rather than just tools. Books like AI Engineering, Machine Learning Design Patterns, and Practical Statistics for Data Scientists are invaluable. They don’t just teach you how to use tools — they teach you how to think critically, solve problems, and build systems that last.

“Music keeps me grounded and creative”

— Whether I’m solving a problem or learning something new, music helps me stay in the zone and unlock my best ideas.

And if you're curious about my favorite movie, Ex Machina is probably my favorite as it touches on the theme of Turing test even though it feels like a low bar today.