DEV Community

Cover image for Frontend Taught Me to Reinvent Myself. AI Is Next
Daniel Ferraz
Daniel Ferraz

Posted on

Frontend Taught Me to Reinvent Myself. AI Is Next

I'm writing a series called "Becoming a Super Engineer": a software engineer since 2013, learning to hold software, data and AI under the same belt, and writing it down as it happens.

Chapter 1 and Chapter 2 were about code. This one is more personal.

If you're a frontend engineer trying to decide whether to stay where you are or move to another role, this one is for you. I don't have all the answers. I do have a direction, and some reasons to be hopeful.

So let me start with a question: have you wondered lately whether there's still room for you?

I have.

Is there still room for us?

I've been a frontend-focused engineer for most of my 13 years in software.

And I'll be honest: some days it feels like the room for that title is getting smaller. Ask an agent for a page with a form and a chart, and you get something decent in a minute.

The first time I saw it, I was impressed. After a while, it started to feel a bit personal.

I think the thing I spent years getting good at is one of the first things AI got good at too.

Most of what I read online didn't help. It was either "nothing will change" or "learn everything, right now", and neither told me what to do on Monday.

If you've felt that, the fear is fair. I won't tell you it's silly.

What helped me was a different question: what did I learn in all these years that has nothing to do with the framework? To answer that, I had to look back.

A dashboard with a form and a bar chart, one bar still being filled in by AI, an "about 1 minute" timer, and an engineer watching

I've switched stacks seven times already

I started in 2013 as an intern, writing vanilla JavaScript and jQuery next to PHP and MySQL. That was my first time writing software for real.

Then came AngularJS, with some Java API fixes on the side. In 2017 it was Angular 2, and Vue right after it.

2018 brought React and TypeScript, D3 for the charts, and my first role as a frontend tech lead. Since 2022 it's been React, D3 and Ionic, leading teams of different sizes, fully remote.

And now, AI and data.

Put like that, it's a lot of starting over. It was.

Every couple of years the ground moved, and each time the choice was the same: learn it, or fall behind. Back then it just felt like keeping up. Looking back, every switch made the next one less scary.

Learning something new, fast, is the skill that never gets old for me. It's also what this AI shift asks from us.

If you've been on the frontend for a while, I bet your list is just as long. Maybe longer :)

Animated timeline: Vanilla JS, jQuery, AngularJS, Angular 2, Vue, React + TypeScript, Ionic, then AI + Data, with one amber line labeled "Learning fast" running under all of them

What stayed when the frameworks left

Two more things stayed with me, and neither of them is technical.

The first is grit. Years of late nights and big delivery dates teach you to finish what you start.

The second is communication: daily contact with clients in another country, and leading teams of different sizes, fully remote. It works the same in any tech you pick up.

Turns out I was never only frontend

The technical side surprised me more.

I spent years on chart and data projects: financial data with market trends, map heat maps, drill-down metrics. You learn what people actually need to see in a pile of numbers, and how to let them dig in without getting lost.

A data platform can hold perfect tables, and somebody still has to turn them into a decision on a screen. Often, a frontend engineer is part of that.

As a junior dev I also shadowed an architect, and I was curious about the whole process. That's where I learned how to model a database and which layers an app needs.

Looking at my path again: PHP, MySQL, Postgres, Java APIs, PL/SQL. The data side was always at the edges of my work. I just never called it that.

It showed up in this series too. Chapter 1 had a React frontend, an API in the middle and governed tables in Unity Catalog underneath, and knowing the layers made the traps between them easier to find.

Chapter 2 was the same instinct, pointed at agents. One misleading line in my tool's output made an agent ignore the right answer. Frontend made me care about what the user actually reads, and with agents, the agent is a user too.

So where am I pointing next?

For most of my career, the next move found me. A new project, a new client, a new framework, and I learned whatever it needed.

This one found me too. This time, I want to excel at it.

Why AI and data, and not backend or management? Honestly, backend was never my go-to. The opportunities came with AI and data first, and they sit right next to what I already liked: charts, data, layers. Management is still an option. It just hasn't happened yet.

So the plan has two steps, in order.

First, the Claude Certified Architect – Foundations exam (CCAR-F). I want to be recognized as a capable AI engineer, and a Foundations exam is the first step on that road for me.

Then the Databricks Certified Generative AI Engineer Associate. That's where I take the AI knowledge and apply it to the data path, and show that what I study turns into something I can use at work.

Why certify, if I already use this every day? I've worked with this knowledge daily for almost a year, and people in my community and some of my colleagues are taking the same exam. For me it's a marker: I studied it, I understood it, and I got tested on it.

I study one hour a day, every day. Weekends too, with longer sessions for full practice tests.

I built the plan with my AI agent. I gave it notes from people who already passed and the official exam guide, and used it to find the patterns. A few came up again and again: only a few courses really matter (2x speed is fine), practice tests matter more than courses as long as they come from several sources, and the real exam is harder than the practice tests. Running out of time is a real risk too.

Then I worked backwards from the goal. My bar is steady practice-test scores before I book the exam, and every full practice test gets a review session after it, so I understand why each answer is right.

It's all over my week: calendar blocks, a checklist, daily reminders and a printed copy on the wall. Hard to forget.

The end of November is my target. If the practice scores aren't there yet, the date moves.

Two cards: CCAR-F in red as step one, with an arrow down to the Databricks Certified Generative AI Engineer Associate in purple as step two, next to a grid of one-hour study days

And where's yours?

My plan won't fit you exactly. You have your own list, and probably more on it than you think.

Grab a piece of paper and give it 15 minutes:

  1. Write down every stack you moved through since you started. Count them.
  2. Next to each one, write what you learned that wasn't the framework. Debugging under pressure, talking to clients, finishing late projects, reading other people's code.
  3. Look at the map below and mark the lines that sound like you.
  4. Pick one small step from there and put a date on it. Mine is an hour a day.

Here's how I see frontend skills mapping to other roles. It's my view, from my own path:

  • Charts, dashboards, drill-downs -> data products and analytics -> build one dashboard on a real dataset, from the table to the screen.
  • Caring about what the user reads -> AI products and agent work -> build one small feature with an agent, and fix the instructions every time it gets lost.
  • App layers and state -> full-stack and backend -> take one feature from the UI all the way down to the database.
  • Communication and leading people -> tech lead and management -> lead the next feature end to end, people included.
  • Learning fast -> any of the above -> pick one line and give it an hour a day.

Three overlapping circles for software, data and AI engineering, with the software circle highlighted and an SE shield where they meet

There's room for frontend engineers, and there's a north

Soooo, is there still room for us?

I believe there is. And I won't pretend this move is like going from Angular to React. This time the tools write part of the code themselves, and that's scary.

But somebody still has to decide what people need to see on that screen, and what the agent needs to read to build it right. I think that's frontend thinking, and it's still needed.

Whether you stay on the frontend or move to another role, you don't start from zero. You start from everything on that piece of paper.

For the Super Engineer path, this chapter is the "why" behind the other two. The software belt I've had since 2013. The AI and data belts I'm working on now.

My next step is the CCAR-F exam. If it goes well, it's my flag on the moon :)

After the exam, pass or not, I'll share the study plan as a cheat sheet you can feed to your own AI agent, along with what I'd change.

Your next step is the piece of paper. 15 minutes, this week.

Let me know in the comments: what did your stack teach you that you're underestimating?

A rocket with the SE shield on its side flying toward the moon along a dashed trail, with a dashed outline of a flag waiting on the moon

Top comments (0)