A tech career in 2026: the fields, the real numbers, and where to start

Same day, same week, I saw both headlines. One said AI was about to wipe out programmers. The other said Brazil is short more than 500,000 tech professionals. They can't both be right at the same time... or can they?

I made the switch into tech coming from chemistry, studying at night after my shift at the lab. So when someone asks me whether it's still worth getting in now, I take it seriously. Yes, it's worth it. That said, there are a few catches that make all the difference.

I'll give you the honest map: what actually changed, where the openings are, with real numbers and sources you can check yourself. No course to sell, no scare tactics.

AI shook up the game, and pretending otherwise would be a lie#

I'll be blunt, because sparing you the unpleasant part wouldn't help at all. AI did shake things up. Mostly in entry-level programming jobs.

Employment for developers aged 22 to 25 dropped nearly 20% from its peak in late 2022, according to Stack Overflow, citing data from the Stanford Digital Economy Lab. And the share of recent grads in hiring at the largest tech companies fell from around 15% before the pandemic to 7%, per the SignalFire State of Tech Talent Report 2025. Demand for juniors softened, and there's no point romanticizing it.

There. I said the ugly part. Now the part the panic headlines conveniently forget.

That drop wasn't the same for everyone. More specialized areas, like security, data engineering and ML infrastructure, were less affected. And AI itself created kinds of work that didn't exist three years ago: model quality, AI evaluation, data curation and AI governance.

What really changed is the kind of junior the market wants. It's not the person who memorizes algorithms to write repetitive code anymore. AI does that, and it does it fast. It's the person who knows how to work with AI: write a good prompt, be suspicious of what it spits out, have enough background to look at the result and say "this part here is wrong." That judgment became the new entry-level edge.

AI didn't lock the door. It raised the step. Whoever shows up knowing how to use the tool gets in ahead of whoever's still terrified of it.

The map: tech is way bigger than "being a programmer"#

This is the part nobody told me when I started, and it would've changed a lot. Look at where the demand actually is.

Software development#

Still the biggest door, but the bar for a pure junior went up. What works well today is someone who pairs code with a specialty (data, cloud, security). Robert Half puts software engineering among the most sought-after roles of 2026 and, at the same time, among the ones hit hardest by a shortage of qualified people: 71% of tech leaders say that gap has delayed projects in the past year. Translation: the jobs exist. What's missing is people who are actually ready.

Data, AI and Machine Learning#

If you're going to bet on a growing area, this is the one. AI-related jobs have grown sharply over the past few years, and "AI Engineer" showed up among the fastest-growing roles in a LinkedIn study on hiring trends, reported by CBS News. And the Bureau of Labor Statistics projects around 34% growth for data scientists from 2024 to 2034 (the closest role to AI and ML in the BLS), way above the average for any profession.

Cybersecurity#

Here the hole is so big it's almost funny. The 2024 ISC2 workforce study estimated a gap of roughly 4.8 million security professionals worldwide, the difference between what companies need and who's actually in the field. To close it, the workforce at the time would have to nearly double.

Just one honest heads-up, because I wish someone had given it to me: according to ISC2, many security managers will consider candidates coming from areas like support, networking and development, even without prior security experience. That's why cybersecurity is usually a great second stop. You get in through support, networking or dev, build a base, and move over.

Cloud, DevOps and Platform Engineering#

It's the invisible backbone of everything that runs on the internet, and almost nobody coming into the field thinks about it. People who know how to build and automate infrastructure are still fought over, and the area is turning into Platform Engineering: teams that build internal platforms so other devs can ship faster. Kubernetes, infrastructure as code, cloud security. All of it pulls salaries up, and a cloud certification (AWS, Azure) really counts on your résumé here.

The doors nobody values (and should)#

Not every path starts with coding. If the junior dev bar scares you, these roles have an easier way in and open a lot of doors:

• IT support / Help Desk. The classic door. You learn how technology actually works in real life and move into networking, cloud or security.
• Data analyst. Less "heavy math" than it sounds. SQL and one BI tool already get you into a job, and it's a natural springboard into data engineering.
• QA / Testing. Knowing how to look at software quality counts for a lot, and AI created demand for people who test and validate systems (AI models included).
• UX/UI and Product. For people who love tech but not necessarily code.

According to Coursera, a good chunk of these roles don't even require a degree and can be reached in months of focused study.

And in Brazil? Here the story is good#

This part genuinely gets me excited. The Brasscom report shows a mismatch that doesn't add up: between 2019 and 2024, the tech sector needed around 665,000 professionals, but the country trained only about 464,000, a 30% shortfall. And the projection points to a cumulative deficit in the hundreds of thousands in the coming years.

And why is that good news for you, specifically? Because most of that gap doesn't come from a lack of people studying. Enrollment in tech courses grew more than 70% between 2019 and 2023, according to Brasscom. But the number of graduates stays far below demand. In practice, a lot of people start the path and never finish it.

About salary, because I know you want to know: the numbers vary a lot (city, company, seniority), so treat them as a ballpark, not a rule. In hubs like São Paulo, a junior dev usually starts in the R$ 3,000 to R$ 5,000 range, according to surveys by Alura and Glassdoor. In more specialized areas like data, cloud and DevOps, a mid-level profile can reach the R$ 8,000 to R$ 15,000 range. And remote work for companies abroad opens up a dollar range, which is a whole other level.

The shift that changed everything, and almost nobody talks about it#

Maybe the most important news for someone starting now isn't which area to pick. It's how companies are hiring.

It went skills-first. According to TestGorilla, around 85% of companies already use skills-based hiring and more than half have dropped the degree requirement. More and more, what counts is what you can do, not where you studied.

In practice, that means three things, and they all work in your favor:

1. Portfolio matters more than résumé. A real project, one you show and explain, is worth more than a list of courses.
2. Technical tests instead of the "degree filter." You prove it by doing.
3. The question stopped being "where did you graduate" and became "what can you build."

I'm not going to sell you that it turned into a walk in the park, though. In practice skills-first is still uneven: plenty of companies say they look at skills, but your résumé goes through an automated filter that cuts by keyword first, and junior openings still get flooded with applicants. When I was trying to break in, I sent plenty of applications that died in silence before anyone looked at a project of mine. It's part of it, and almost nobody talks about it.

I got in myself without a CS degree. I'm finishing software engineering now, but the door opened because of what I knew how to do. The game rewards the people who build.

Where to start (the plan I'd give my 2023 self)#

Put it all together and the path shows up. There's no magic. There's consistency.

1. Pick ONE area and embrace it. The number one beginner mistake is jumping from topic to topic. Look at the map up there, see what genuinely makes you curious (you're going to spend hours on it) and focus. You can change later. What you can't do is chase five things at once.
2. Fundamentals before the trendy framework. Logic, a bit of how the internet works, the base of your area. Fundamentals stick around; the trendy tool you'll swap out in a couple of years.
3. Learn by doing. Watching videos isn't learning. It's watching someone else learn. Build a small, real project from week one. That's where your portfolio comes from.
4. Use AI as a tutor, with suspicion. Ask it to explain things, generate examples, review your code. But always double-check. Knowing how to work well with AI is, today, a real edge when companies hire.
5. Think about an entry-level certification. CompTIA, Google, AWS, Microsoft. They help prove your level, especially in support, cloud and data.
6. Be patient with the timeline. It's not weeks: it's months of focused study until the first job, and the timeline varies a lot from person to person (your background, your pace, the market). In the middle it'll feel slow. That's normal. The person who keeps going when it gets boring is exactly the one who joins the 5% who finish.

To wrap up#

The 2026 market isn't the easy paradise of a few years ago, and it isn't the desert that AI fear paints either. It's demanding and short on people at the same time. The bar went up, but the demand for people who know how to do the work has never been bigger. In Brazil, hundreds of thousands of professionals are missing.

If I could leave you with one single idea, it'd be this: most of your competition will quit halfway. The 2026 barrier isn't AI, and it isn't a lack of jobs. It's persistence.

I traded the lab bench for VS Code in my early twenties, from scratch, juggling work and study. If I managed to start over, you can too. Pick a path, build real things, use AI in your favor and keep going when it gets hard. The market is waiting for you for exactly that reason.

If you want to talk through where to start, hit me up on LinkedIn. Seriously, reach out.