Are Junior Devs Going Extinct?

Will AI coding agents replace all junior engineering roles?

A surge of engineering leaders are sounding the alarm over how AI is gutting junior engineering roles. Will this dramatically destabilize engineering orgs?

I asked our tech leadership team what they thought. Here’s what they said.

Why We Still Need Junior Engineers in the AI Era

Don't stop hiring juniors. Change how you train them, or you'll lose the human authority needed to validate what AI produces.

Hari Chandrasekhar

SVP of Engineering, Core

We are facing a long-term risk if we replace juniors with AI. The solution isn’t to stop hiring juniors, but to completely change how they are trained.

AI is here to stay, and it’s naturally going to take up the low-level repetitive tasks that juniors traditionally used to start on. But instead of treating this as the new norm and completely stop recruiting junior engineers, engineering orgs need to shift the junior playbook. 

We need to teach them to think bigger, focusing on design, architecture, and system-level problem solving right out of the gate.

It also doesn’t mean they can skip the fundamentals. If junior engineers don’t understand the foundation of a system, we will lose the human authority required to validate and control what AI spits out.

Optimizing for short-term productivity by replacing younger engineers with AI will become a massive trap that could impact long-term technical depth. The focus should be to use AI as a productivity improvement tool, and not primarily as a cost cutting technique.

AI Is Reorganizing Engineering, Not Replacing It

Junior engineers will matter even more in the AI era. The standards are just changing.

Benji Thomas

Co-Founder & CTO

The core theme here is right. AI is giving senior engineers a real boost. But the fear around junior engineering collapsing feels overblown to me. We’re in the middle of a software industrial revolution, and like every tech revolution before it, the work gets reorganized, not erased.

Everyone’s focused on code generation right now, but testing, validation, observability, and regression detection are going to catch up fast. Probably in the next 6 to 9 months.

Engineering is simply changing shape; companies have always over-hired for repetitive implementation and maintenance toil, and AI is transforming that layer.

Good junior engineers will still matter, maybe even more, the standards just changed. It will be less about cranking out CRUD code, and more about systems thinking, debugging, product intuition, architecture, and knowing how to orchestrate AI.

Apprenticeship isn’t going away, but it won’t look like “spend 3 years fixing tickets” anymore. Juniors will probably learn faster by running AI-assisted systems than by hand-coding everything from scratch.

The instability we’re seeing isn’t engineering dying. It’s a just big reorganization of how the work gets done, and the world is about to move a lot faster.

AI Is Redefining What It Means to Be a Programmer

The role of the programmer will shift into deciding what to build, not how.

Nikhil Gopinath Kurup

SVP of Engineering, ML Platforms

This made me think of the quote, "People tend to overestimate what can be done in one year, and underestimate what can be done in five or ten years"

And I think that is what is happening here. 

This kind of technical revolution is nothing new. Back when computer programming used to be done in binary, and people started building compilers, there was pushback from people who did not like these "mad revolutionaries."

The original definition of a computer programmer was someone who actually took handwritten code and punched it into a machine. That has been replaced by programmers who actually write code in high-level languages. 

In my opinion, this kind of change is exactly what is going to happen next: the definition of a programmer is going to change. 

In the future, converting requirements into code is not going to be an important part of the value chain. A lot of that will be replaced by tools and automated processes, just like CI/CD tools that compile and deploy.

The most valuable part of this entire picture is actually identifying what needs to be built. Building is easy, deciding what to build and what actually gets value is going to be the harder part of the picture. And that is what the new role of a programmer will be.

Given this hypothesis, I would suggest we should start training these future programmers by becoming better product managers, better business analysts, and better program managers, so that they can go broader rather than just deeper on the technical side.

Running Engineering With Agents Is Possible

If self-managing systems like Ramp's keep proving out, the engineering team of the future may look nothing like what we're used to.

Shankar Jothi

VP of Engineering, ML

AI is increasing the pressure to move faster, ship more, and do more with leaner teams. As a result, investing heavily in training junior talent may not always feel like the immediate priority.

There are interesting shifts happening in how teams operate. Ramp, for example, is building a sustainable model where systems are self-managing, from identifying issues to fixing them autonomously. In my own experience, AI tools can already make certain implementation and operational tasks significantly more reliable and efficient.

It also raises a broader question: if agents continue improving, how much of the engineering workflow (including higher-level coordination and decision support) could eventually become more automated?

That said, I don't think this necessarily means junior engineers become unimportant. Long-term innovation still depends on developing future technical leaders, strong product thinking, and people who can navigate ambiguity beyond what current systems can handle.

Maybe the bigger shift is that the structure of teams changes. Smaller, highly experienced teams equipped with strong AI tooling may be able to accomplish much more than before, while companies continue balancing short-term efficiency with long-term talent development.

My current takeaway is less "AI replaces engineering teams" and more: AI is fundamentally changing what engineering teams may look like in the future.

Give your teams more time back so they can build the future of engineering. See Sedai in action.