My inbox tells a clear story. Every single request I get for an engineer is for a senior profile. Not just senior, often very senior: staff level, principal, someone who can lead a complex project from day one. I run Amplify IT, and I connect Brazilian engineers with US companies. My job is to find the right person for the right role. I can't remember the last time one of them was for a junior.
The market is not asking for juniors anymore
This isn't about budget, at least not in the conversations I have. It's about demand. The market simply isn't asking for juniors like it used to. When I talk to engineering leaders, they rarely mention building out a pipeline of new talent. They need someone who can hit the ground running, someone who's seen a specific problem before and knows how to fix it. They need a senior.
What turns a competent engineer into a senior one? It's not just more years in a chair. It's the accumulation of specific experiences. It’s making a bad call on an architectural decision, seeing the consequences in production, and then having to untangle the mess. It’s shipping code that breaks, understanding why it broke, and implementing a more resilient solution. It’s being wrong, finding out you’re wrong, and fixing it. That loop of mistake, diagnosis, and correction is the crucible of seniority.
AI is changing how engineers learn and grow
Now, think about that loop today. Specifically, think about how AI tools fit into it. A junior engineer, fresh out of school, might have spent weeks wrestling with a tricky bug or trying to optimize a slow query. They'd read documentation, dig through logs, ask for help, and eventually, after a lot of effort, figure it out. That struggle, that debugging process, that's where the deep learning happens. That’s where the mental models form.
Today, a junior might ask a code assistant for the solution. The assistant provides it. The bug is fixed, the query is optimized. The immediate problem is solved. But what did the junior engineer actually learn? Did they understand the underlying system deeply? Did they internalize the nuances of the database engine or the network protocol? Or did they just get a correct answer without the struggle that builds true understanding?
The data is starting to show this. Fastly surveyed 791 US developers in July 2025. They found that 32 percent of senior developers said more than half the code they ship is AI-generated. For juniors, that number was 13 percent. That's a significant difference in reliance. The survey also found that just under 30 percent of seniors said they edit that AI output enough to offset most of the time it saved them, against 17 percent of juniors. There's an obvious objection to reading much into that, and it's a fair one: seniors use AI more, so more of them would run into the editing problem no matter what. These two numbers on their own don't settle it. But the read I'd bet on is that seniors are catching things juniors aren't yet, because catching them is the part that takes years. (It's worth noting this was a self-reported survey covering US developers only, and Fastly sells CDN and edge infrastructure, not code quality tooling.)
The senior engineer, even when using AI, still engages in that critical loop. They use AI to generate a first pass, then they tear it apart, optimize it, make it fit their specific context, and learn from what the AI missed. The junior might just accept the output, close the ticket, and move on. The loop of being wrong and fixing it is getting shorter, or in many cases, it’s being closed by a model before a junior engineer even fully enters it.
The long-term cost of this shift is unclear
Nobody feels this problem today. This quarter, your senior engineers are productive. They're using AI to accelerate their work, shipping more code, solving more problems. It looks like a win. The demand for more senior engineers continues because they deliver value. You don't need to hire juniors because the existing team is doing more with less.
But what happens in three, five, or seven years? Where do the next generation of senior engineers come from? If the mechanisms for accumulating experience—making mistakes, debugging complex systems, deeply understanding failures—are being short-circuited by AI for those early in their careers, then the pipeline of future senior talent is quietly drying up.
The cost of a senior engineer will only continue to rise, driven by scarcity. The supply won't keep up with the demand. I don't have a solution for this. I place senior engineers, so in the short term, this trend is actually good for my business. But it's a structural shift that will create real problems for engineering leadership down the line.
We're building teams that rely heavily on experienced people, without actively cultivating the conditions for new experienced people to emerge. The engineering leaders I talk to are focused on immediate needs, on delivering now. The long-term implications of where real seniority comes from, and how it’s built, are not on their radar today. But they will be.
