I'm Martin Vrkoč. I primarily launch my own projects. And when a project enters the mode of "this should have been done yesterday," that's when I step in: I bring direction, structure, and pace. I help companies with software project management, analysis, and development, today mainly through AI (LLM), which I can turn from hype into real team performance. And outside projects? Running, skiing, and golf, because that's where you also see who can handle pressure and finish strong.
Vibe-coding is a good start. Production software needs specification, tests, security, and iteration. This is how modern development becomes predictable, operable, and safe.
Most companies use AI as a better search interface. The biggest value appears when AI runs 24/7, gathers data, prepares offers, and moves sales forward without waiting for a click.
Companies often assign AI what is hardest even for humans. Results are weak. A better strategy is the opposite: move 80% routine to AI and keep 20% creative decision-making with people.
AI can speed up work, but without the right approach it often creates only the impression of higher productivity. What works in practice and what does not?
