How to Build a Career in AI from dev to Director of AI: The 5 Biggest Problems
Dalibor Jovic 
If you’re considering AI as your career path, I’m going to show you how I built my career around it, and the five biggest problems
If you’re considering AI as your career path, I’m going to show you how I built my career around it, and the five biggest problems that can cost you months (or years) if you don’t see them coming.
This matters because AI isn’t just “another tech trend” you casually add on top of your job. Done seriously, it becomes an identity shift and a career accelerator, but only if you learn how to handle the real obstacles: the speed of change, internal politics, resistant coworkers, your own motivation, and the skill gaps nobody warns you about.
If you learn these five lessons early, you’ll move faster, earn trust sooner, and avoid the painful loop of “I’ll start once I feel ready.” Because in AI, that day never comes.
Before the five problems, a quick piece of context: I coded for 18 years. I cared deeply about clean architecture, SOLID principles, and best practices. I assumed I’d do that forever.
Then I started spending more time in boardrooms as a senior consultant, and I saw the business side more clearly: companies don’t pay for code. They pay for solutions.
At the same time, I got involved in AI adoption internally, helping hundreds of colleagues use AI properly. And it became painfully obvious: the thing I’d spent 18 years mastering was exactly what AI was becoming exceptionally good at.
AI can write code faster. Even when it’s wrong, you ask again and it fixes it, often still faster than a human would.
For a while, it felt like my career was falling apart. Then it clicked:
The value was never the code. The value was bridging the problem and the solution.
So I stopped centering my identity on “the person who writes the code,” and I moved toward AI adoption and leadership. Today I work as a Director of AI, and I also run one of the largest AI communities in my city (300+ members, recently supported by Google). And here are the five problems I had to overcome to get here.
AI Moves Too Fast (So Stop Trying to Learn Everything)
Why this matters: the pace of AI change will exhaust you if you approach it like normal tech. You can spend weeks learning a model, a tool, or a concept, and the moment you feel on top of it, everyone’s talking about a “revolutionary” new thing (usually announced on a Tuesday).
If you don’t handle this, you’ll live in permanent FOMO, constantly behind, constantly frustrated, never feeling credible enough to step up.
What it looks like in real life:
- Without a strategy: you chase tools, burn out, and end up stuck as “AI enthusiast.”
- With a strategy: you stay current without losing your sanity, and you start building real authority.
What worked for me (two parts):
1) Delegate learning by being around other AI-minded people.
Physically or digitally, doesn’t matter. What matters is that you don’t try to absorb everything alone. Have colleagues (or even friends) explore specific tools and report back. Organize knowledge sharing. Ask people to present what they learned, and share what you learned too.
This is exactly why I started a local meetup, so the learning load is shared. We’ve had people from Google, the police, and large companies share research and practical experience.
2) Be systematic, make learning sustainable.
You don’t win AI by reading headlines. You win by building a learning engine you can run for years.
And here’s the funny part: after enough exposure, you start noticing a pattern, most “new tools” are variations of the same underlying ideas (context management, memory, workflow wrappers). The hype changes faster than the fundamentals.
Executives Say They Want AI, But Often Won’t Enable It
Why this matters: people above you can make AI adoption easy… or impossible. Many leaders will say they want “more AI,” but then expect:
- 8 hours of regular work plus 8 hours of AI learning
- experimentation without paid tools or licenses
- progress without time, budget, or risk tolerance
If you can’t manage up, you’ll be stuck. If you learn to manage up, AI becomes one of the easiest chances you’ll ever get to jump multiple career steps.
What it looks like in real life:
- Without managing up: you talk about models and features; they talk about budgets and risk; nothing happens.
- With managing up: doors open, because leaders see you can help the business, not just “play with tools.”
What worked for me: learn to speak business possibilities and business consequences.
That means showing up prepared and learning the basics executives care about:
- Tradeoffs between AI strategies (and which fits which type of company)
- Risk levels and what “responsible” adoption actually means
- Clear upside: money saved, money generated, time reduced
- Real costs: licenses, token usage, salaries, operational overhead
- The difference between an expense and an investment
- What “strategy” even means, and having a few concrete options
When you can explain AI in business terms (not buzzwords), you become valuable in a different way. You’re no longer “asking for an AI license.” You’re helping them make better decisions.
Your Peers May Resist AI (Because It Scares Them)
Why this matters: even if leadership is on board, you can’t drive adoption alone. If the people around you don’t believe in it, or feel threatened by it, your progress slows to a crawl.
And the pushback can drain your energy fast:
- “AI can’t replace humans.”
- “It forgot what I said earlier.”
- “It hallucinated once, so it’s useless.”
- “It’s a bubble.”
You’ll be focused on what AI can help with. Others will be focused on what it can’t do. If you take that personally, you’ll burn out.
What it looks like in real life:
- Without patience: you push harder, meet more resistance, and become “that AI person” nobody wants to talk to.
- With patience: you slowly create safety, and people begin experimenting without feeling stupid or threatened.
What worked for me: assume the resistance is coming from one place, fear and insecurity.
So instead of arguing, I tried to teach:
- Ask people to explain why they think something won’t work
- Explore how the failure happened (and how to avoid it next time)
- Challenge them to find one task AI can speed up safely
- Stay patient, because for some colleagues the shift took a year or more
This is slow, unglamorous work. But it’s how you turn “AI is a threat” into “AI is a tool we control.”
The Hardest Battle Is Your Own Motivation (and Identity)
Why this matters: you will hit the wall, repeatedly. Not because you’re not smart enough, but because the environment will test you.
People will support you privately and then disagree publicly. Someone will promise resources and then change their mind. You’ll experience moments that feel like betrayal, even if nobody meant it personally.
I’ve been there. One time it hit so hard I almost quit my AI path and started going back into other languages, r
eady to throw away three years of work.
What it looks like in real life:
- Without a clear “why”: setbacks feel like proof you should stop.
- With a clear “why”: setbacks are painful, but they don’t derail you.
What worked for me: get honest about your reasons, and the price you’re willing to pay.
Ask yourself:
- Why is AI important to you?
- What are you willing to change to pursue it?
- Does this require changing what you do daily?
- Does it mean shifting your identity?
- Does it mean changing teams, or even companies?
For me, the identity shift was real. Closing the IDE felt like betrayal at first. Changing my LinkedIn from “Frontend” to “AI adoption” felt like stepping into the unknown.
But I also wanted more business impact and responsibility. And eventually, to get the opportunity I wanted, I did change companies. I’m not saying you should do that, but I am saying: your motivation has to come from you, because you’ll need it when things get hard.
Think long and hard about your “why.” You’ll need it more than once.
You’ll Feel Underqualified, So Build Skills Systematically
Why this matters: once you become “the AI person,” you’ll get pulled into rooms with seniors, executives, and very smart people. They’ll ask questions like:
- “What should our AI strategy be?”
- “Do we need to hire someone?”
- “What should we do next Monday, and the week after?”
If you’re not prepared, you’ll feel exposed. Not because you lack intelligence, because nobody taught you the surrounding skills: strategy, change management, communication, prioritization.
What it looks like in real life:
- Without structured learning: you react, ramble, over-joke, under-joke, miss the point, and lose trust.
- With structured learning: you become concise, useful, and calm, even under pressure.
What worked for me: treat learning like a system, not a mood.
Block time in your calendar, one hou
r every second day (or daily). And don’t just consume random AI news. Pick one skill and go deep:
- One month: strategy fundamentals
- Next month: change management
- Next: presentation skills
- Then: facilitation, stakeholder management, and so on
If you truly want a serious AI role, you eventually make a deal with yourself: this isn’t a strict 9-to-5 transformation. For a period of time, it becomes a calling. The “AI person” doesn’t stop at 5 PM, because the identity shift requires reps.
Finally, what I hope for you
I hope you take this path with your eyes open.
Not everyone needs to quit engineering, if you love coding, keep it. But if you want AI to be your career, understand what you’re really signing up for:
You’ll have to keep up with fast-moving technology without burning out.
You’ll have to learn to influence executives.
You’ll have to bring peers along without becoming bitter.
You’ll have to protect your motivation through setbacks.
And you’ll have to build missing skills systematically, until you become the person who can bridge problems to solutions at a higher level.
That’s the real work. And if you do it, the payoff isn’t just a new job title, it’s becoming someone who can shape what happens next.
I’m wishing you a strong, patient, disciplined journey forward.
Categories: : AI
