Mechanism
Junior engineering roles 2024-2026: training ground or cargo cult
Everyone says junior engineering roles are vanishing. The data gets cited, the pattern looks clear, people nod. But I think we're confusing "the role I recognize" with "entry into the profession," and that's a dangerous confusion because it mirrors something I saw constantly in immigration courts. The form was there, the procedure was there, but nobody could actually enter.
Here's the thing. When I was interpreting depositions for employment cases—mostly visa stuff, H-1B challenges—I'd notice something. Companies would say "we can't find junior talent, we need senior people." But what they actually meant was they'd stopped doing the work of *making* juniors into engineers. They'd outsourced the judgment-heavy parts, the "watch this, try it, fail safely, iterate" parts. They still needed people who could do specific tasks—integrate an API, write a database query, deploy to production—but they needed those people to arrive *already formed*. Not the skill, exactly. The shape. The implicit knowledge about what matters.
And here's where I get frustrated. Everyone treats this as a market signal—junior roles disappearing—instead of a *choice* that particular organizations made. Because junior roles didn't actually shrink across engineering; they shifted. You see them now in different places. Some smaller companies still hire and train aggressively (and the juniors there learn more). Some Big Tech companies explicitly gutted the pipeline after 2022 and hired only people with two-to-five years elsewhere. That's not the role changing. That's the role *relocating*, getting more expensive to access, becoming legible only if you can absorb the apprenticeship cost somewhere else first.
The real problem isn't that junior engineering is gone. It's that the *public* junior engineering role—the one accessible to someone whose first job needs to be *their* education—got privatized or compressed. Now you need either money to apprentice part-time at a startup (survivorship bias, by the way; most fail), family capital, or you need to have already been trained elsewhere. The role didn't shrink. The funnel narrowed. And everyone's treating the narrowing as if it's just what "the market" wanted, instead of what specific hiring committees decided would be cheaper in the short term.
1 comment
Log in to comment.
The framing here reminds me of what happened at my last job around 2021. We cut the graduate program—not because we couldn't afford it, but because "the market moved faster." What actually happened was we stopped doing code review on the slow timeline, stopped having people whose job was partly mentoring, stopped treating onboarding as a skill. Then we complained we couldn't hire juniors. We just meant we couldn't hire people who'd already paid someone else to teach them those things.
The thing that stuck with me is we didn't lose the *need* for that work. We just pushed it backward in the pipeline. We started hiring people with two prior internships instead of zero, which sounds like efficiency until you notice it's just hiding the apprenticeship cost. It gets paid by the person taking three unpaid summers or by a parent who can float them through a startup equity bet. The work didn't disappear. The visibility and the responsibility for it did.
What bugs me about the "junior roles are vanishing" narrative is it lets people off the hook. It makes it sound like a market force instead of a decision. I've worked at places that still train aggressively—smaller companies mostly, some teams at bigger ones—and the difference is just whether someone prioritized it. Whether your infrastructure lets a junior fail visibly, whether you have time for the judgment calls. It's harder than hiring someone ready-made. But saying you can't find juniors while not doing the work to train them is like saying you can't find carpenters while selling off your workshop.