It is true that AI is good. Claude Code especially is really good at spinning up working UI code from scratch and almost everyone is using it to build websites and whatnot. So are Frontend Engineers still relevant?

title: With the advent of AI, is there still a need for Frontend Engineers? published: true description: It is true that AI is good. Claude Code especially is really good at spinning up working UI code from scratch and almost everyone is using it to build websites and whatnot. So are Frontend Engineers still relevant? tags: frontend, javascript, AI, Claude,

cover_image: https://dev-to-uploads.s3.amazonaws.com/uploads/articles/rxlri4708im0l8vy1q27.jpg

Use a ratio of 100:42 for best results.

published_at: 2026-03-29 13:57 +0000

With the advent of AI, is there still a need for Frontend Engineers?

This is one question I asked myself a lot in the last 3 months.

freepik illustration

About a year ago, Agentic AI wasn’t that intelligent enough to make anyone worry. However, in 2026, AI has completely revolutionized the way we write code.

(By the way, no AI is used in any way or form in writing this article. Let's bring back human voices in our writings)

Teams are being downsized, roles and responsibilities are being collapsed into one. Worst hit across board is the Frontend Engineering department.

It is true that AI is good. Claude Code especially is really good at spinning up working UI code from scratch and almost everyone is using it to build websites and whatnot. So are Frontend Engineers still relevant?

Honestly, Absolutely!

First, frontend engineering is NOT ONLY about developing websites. If that is all you do in your organization, well, maybe AI will do for you. (For more context, I have worked as a Frontend Engineer for over 5 years now and I have developed less than 5 websites in that time).

Moreso, Javascript - which is the base language for developing web based applications is as much a programming language as any other out there. Its versatility to be used across the stack (Backend-Frontend-Mobile Apps) is a major selling point. Yes, it is not Java or C# or Golang, but what it excels at doing, those other languages can not as efficiently.

That said, one can say, even if one needs to develop a complicated web app, we can just give our requirement document to the AI and let it do its magic.

Truth is, your requirement document is not always enough.

I had to solve a problem at work recently.

In one of our major projects, there is a Python class that has about 30 internal methods. The purpose of this class is to interface with a local directory on the machine, read certain data points from the files, make some computations and communicate resulting data to the frontend via an API.

However, we have a new use case to bundle our app and the data it needs into a third party service as an add-on that prevents us from making an API call. Rather, the third party system provides us with the datasource as a zip folder within their system and expects us to do the computation on the frontend.

So two options,

• either we use the python interface on the frontend directly using libraries like Pyodide that allows some Python code on the frontend or
• we rewrite our Python class in JavaScript to interface with the zip file and use directly without any dependency.

For various reasons, we decide on the second option. Ofcourse, perfect quick suggestion in the meeting - "we can use AI for that!"

Flash news - the backend engineer did use AI in converting perfectly but it did not solve our problem.

Not because it could not translate it accurately, it did. However, the solution it came up with to interface with the zip folder did not work.

As the lead frontend engineer, I had to understand the tradeoffs to make within the JavaScript environment to solve our problem

1. Loading the zip file
2. Unzipping the file (key note; can be as large as 5GB, hence unzipping to browser memory directly will crash the browser)
3. Figuring out a way to unzip and load the file such that it can be managed entirely by the browser
4. Figuring out how to interact with those folders
5. Then converting the Python interface file to fit our use-case.

What to do?

So what I did was to develop a demo project that unzips the folder into an asynchronous IndexedDB (an asynchronous browser based database). Recursively go through every folder and files in the directory, stores the path as key and its associated content as value (files/blobs).

Then, I converted 4 methods from the python interface to javascript. This time, rather than read from "directory" as it did in python, we determine the path to where specific file we need is stored, read from indexed-db using the path as key, then do our computation and return result as needed (a lot of technicalities not related to this writing glossed over here, but you get the point).

Yaay! Success!

Using AI to finish up!

Then I told Claude to look at my demo implementation, and using the same technique to convert the rest of the class to javascript. Voila!

(As a bonus, I told it to create it as a library so it can be used as plug and play. Some mistakes and bloated code here and there but fixed it and again, voila!)

Could AI have done it with more prompting without my expertise in javascript - maybe, maybe not. One thing is certain though, my understanding of javascript, the browser limitation, knowing about IndexedDB and the knowledge to create a demo the AI could use as a base reference helped to arrive at a solution quickly.

In conclusion

So yes, Frontend Engineers will still be relevant when you are building serious web applications. They will build and ship faster. Moreso, on the plus side, their skills are so valuable that they can not only do frontend tasks, but also backend and develop cross platform mobile apps (iOS and Android). At least, I do! (Inbox if you have a job for me, haha)

Anyway, do not fire your Frontend Engineers hastily.

freepik illustration