The 67% Junior Hiring Collapse: Are We Creating a Hollowed-Out Career Ladder?

[eng_director_luis]

I mentor through SHPE (Society of Hispanic Professional Engineers), and what I’m hearing from computer science graduates right now is genuinely alarming. Three of my mentees graduated last spring with solid CS degrees, internship experience, and strong portfolios. Two of them are still looking for work. Not because they’re not talented — because the jobs they were trained for are disappearing.

The data backs up what I’m seeing on the ground: U.S. entry-level tech job postings dropped 67% between 2023 and 2024, according to Stanford’s Digital Economy Lab. In the UK, it’s 46% and projected to hit 53% by end of 2026. Among the 15 biggest tech companies, recent graduates went from 15% of all hires pre-pandemic to just 7% today.

This isn’t a market correction. This is a structural shift that threatens the entire talent pipeline of our industry.

The Harvard Study That Should Scare Every Engineering Leader

A new Harvard study examined 285,000 U.S. firms and 62 million workers over a decade. The findings:

• When companies adopt generative AI, junior employment drops 9-10% within six quarters
• AI-adopting companies hired five fewer junior workers per quarter than before
• Senior employment barely changed

The critical detail: this wasn’t layoffs — it was hiring freezes. Companies didn’t fire juniors; they simply stopped replacing them. The jobs vanished quietly, without headlines.

54% of engineering leaders in a 2025 LeadDev survey said they plan to hire fewer juniors thanks to AI copilots. Salesforce’s Marc Benioff announced “no new engineers in 2025.” The message from leadership across the industry is clear: AI tools let seniors do more, so we need fewer juniors.

Why This Is Short-Sighted

Here’s the argument I keep making to my own leadership team, and I’ll make it here too:

Today’s juniors are tomorrow’s senior engineers.

A 67% hiring cliff in 2024-2026 means 67% fewer potential engineering leaders in 2031-2036. There’s a 70% likelihood of a crisis point in 2029-2031 when mass senior retirements collide with a talent shortage we created ourselves.

The U.S. already faces a projected shortage of 1.2 million software engineers. We’re making it worse by shutting the front door.

The “hollowed-out career ladder”

We’re creating an industry with plenty of senior engineers at the top, AI handling routine tasks at the bottom, and nobody learning the craft in the middle. The tasks that used to train juniors — writing unit tests, fixing CSS bugs, building CRUD endpoints, documenting APIs — are exactly the tasks AI now handles.

Think about it: the way senior engineers became senior was by doing tedious, repetitive work for years. That’s how you build intuition. That’s how you learn to debug. That’s how you develop the judgment to know when AI-generated code is subtly wrong. If we automate away the training ground, who reviews the AI’s work in 10 years?

AI isn’t even the real driver

AI became a convenient narrative. “AI made juniors obsolete” sounds better in board meetings than “we can’t afford to train anyone in this economic climate.” If AI truly made juniors obsolete, the collapse would have started in late 2022 when ChatGPT launched. Instead, it accelerated in 2023-2024 when interest rates spiked. Companies stopped investing in training because capital got expensive, and AI gave them cover to justify it.

There’s also a surplus of laid-off mid-level engineers competing for anything labeled “junior.” The market is flooded with experienced talent willing to take a pay cut, which further pushes actual juniors out.

What I’m Doing About It

I lead 40+ engineers, and I’ve pushed back against the “no juniors” trend:

We still hire juniors. Two this year. My argument to leadership: the cost of training a junior for 12 months is less than the cost of a bad senior hire who doesn’t fit the culture, and the loyalty you build is worth more than the short-term efficiency loss.

We’ve redesigned the junior role. Our juniors don’t do the old grunt work — AI handles that. Instead, they:

• Review AI-generated code (learning to spot quality issues)
• Write integration and end-to-end tests (higher-order thinking than unit tests)
• Pair with senior engineers on system design (exposure to architectural decisions)
• Rotate through teams every 3 months (breadth of experience)

We’ve created a “medical residency” model. Inspired by the medical training pipeline, our juniors spend their first 6 months in a supervised sandbox environment. They fix broken AI-generated code, debug intentionally flawed systems, and build features in a safe-to-fail environment before touching production.

We track mentoring hours as a team metric. Senior engineers get credit in performance reviews for time spent mentoring juniors. If we don’t incentivize knowledge transfer, it doesn’t happen.

The Industry Needs to Act

AWS CEO Matt Garman called replacing junior developers with AI “the dumbest thing.” He’s right. We’re eating our seed corn.

The companies that keep hiring and training juniors through this AI transition will have the strongest engineering organizations in 2030. The companies that cut their pipeline will be scrambling to hire the same seniors everyone else wants — at premium prices — because nobody invested in growing talent.

I’ll say it directly: if your company has zero junior engineers, you’re not running a sustainable engineering organization. You’re running a contractor shop with better benefits.

How is your company handling junior hiring? Are you still investing in the pipeline, or have you joined the freeze?

[vp_eng_keisha]

Luis, this thread is deeply personal for me. I’m a first-generation Black woman in tech leadership, and every rung of my career ladder existed because someone invested in me when I was junior. If that ladder had been hollowed out when I graduated from Spelman, I’m not sure I’d be a VP of Engineering today.

The Diversity Pipeline Is Getting Hit Hardest

Here’s what the aggregate data doesn’t show: the junior hiring collapse is disproportionately affecting underrepresented groups. The companies most aggressively cutting junior roles are the same large tech companies that had the most robust diversity hiring pipelines.

At our EdTech startup, I’ve been tracking this closely:

• The number of HBCU recruiting events from major tech companies dropped by more than half between 2023 and 2025
• Bootcamp graduates, who skew more diverse than CS degree holders, are facing near-zero hiring rates at large companies
• Internship-to-full-time conversion programs — the #1 pipeline for diverse junior talent — are being cut or frozen

When you combine the 67% junior hiring decline with the fact that diverse candidates are disproportionately represented in junior cohorts, the DEI implications are devastating. We spent a decade building pipelines to bring more women, Black, Hispanic, and first-generation engineers into tech. Those pipelines are being shut off.

Why I Still Hire Juniors (Despite the Pressure)

I’m scaling from 25 to 80+ engineers. My CFO has asked me, multiple times, why I’m hiring juniors when I could hire two fewer seniors who “produce more.” My answer is always the same:

1. Culture. Juniors ask “why?” in ways that challenge assumptions seniors have internalized. They keep the team honest about complexity, documentation, and knowledge sharing.

2. Retention. Engineers we grow from junior to mid-level have 3x the retention rate of senior external hires. They know our systems, our culture, and our customers in ways no senior hire can replicate in their first year.

3. Cost sustainability. An all-senior team is expensive and fragile. If two seniors leave, you’re in crisis. A healthy mix of levels means you have a bench, and your seniors have force multipliers instead of doing everything themselves.

4. Moral obligation. I benefited from programs that invested in my growth. If I pull the ladder up behind me, I’m part of the problem.

The “AI-Free Weeks” Update

In the AI sentiment thread, I mentioned mandating AI-free weeks for junior engineers. I want to share what we’ve learned after 3 months of running this:

• Juniors who did AI-free rotations debug 40% faster than those who didn’t. They’ve built the mental model of how code works, not just how to prompt for it.
• The AI-free juniors report higher confidence in code reviews. They trust their own judgment more because they’ve practiced forming it.
• Senior engineers report that AI-free juniors ask better questions. Instead of “why doesn’t this AI-generated code work?” they ask “I think the issue is X because of Y — can you confirm?”

The medical residency model Luis describes is exactly right. You can’t learn surgery by watching an AI perform it. You have to hold the scalpel.

What We Owe the Next Generation

I’ll be blunt: if you’re a senior engineering leader who benefited from mentorship and training early in your career, and you’re now cutting junior hiring because AI copilots let your seniors type faster, you’re being a hypocrite.

We didn’t get here on our own. Nobody becomes a senior engineer in a vacuum. The industry owes it to the next generation to keep the door open — and to redesign the training, not eliminate it.

[alex_dev]

This one hits close to home because I was one of those juniors not that long ago. I got my first role in 2019, right before the pandemic. If I were graduating today with the exact same resume, exact same skills, exact same hustle — I genuinely don’t know if I’d get hired.

The Door That Closed Behind Me

My first job was at a 40-person startup. They hired me because I showed enthusiasm, could build a basic React app, and passed a take-home project. My first six months were rough — I broke staging twice, wrote code that my senior had to essentially rewrite, and asked questions that probably made people wonder if I’d ever figure it out.

But I did figure it out. By month 8, I was shipping features independently. By year 2, I was mentoring the next junior. That pipeline worked.

Today, that same startup has 120 engineers and hasn’t hired a single junior in two years. Their reasoning? “We can’t afford the ramp-up time when AI can fill the gap.” Except AI isn’t filling the gap — it’s generating code that their seniors spend time reviewing and fixing, which is exactly what they used to do with juniors, except now there’s no human learning on the other end.

The Numbers Don’t Lie, But They Don’t Tell the Whole Story

@eng_director_luis, your point about interest rates being the real driver resonates. I was talking to a recruiter friend last month, and she said something that stuck with me: “Companies aren’t replacing juniors with AI. They’re using AI as the excuse to not backfill positions they cut during the rate hikes.”

The timeline supports this. Entry-level postings started collapsing in mid-2023 — months after the Fed rate hikes, not months after ChatGPT launched. The AI narrative is convenient because it sounds innovative rather than cost-cutting.

But here’s where it gets complicated: even if AI isn’t the cause, it’s becoming the justification for making the cuts permanent. “We don’t need juniors because AI” is an easier board presentation than “we don’t want to invest in talent development.”

What I Tell New Grads (And It Feels Hollow)

I mentor three CS students through a local bootcamp partnership. The advice I gave two years ago — build projects, contribute to open source, network, apply broadly — feels increasingly disconnected from reality when they’re competing against 200+ applicants for roles that now require “3-5 years experience” for what used to be entry-level positions.

The most honest advice I can give now:

1. Learn to work with AI, not just code — The juniors who will get hired are the ones who can demonstrate AI-augmented productivity from day one
2. Target companies under 50 employees — Smaller companies still need humans who can wear multiple hats; they can’t survive on AI-generated code alone
3. Build things that ship, not just portfolio projects — Deploy to real users, even if it’s 10 people. The gap between “I built a todo app” and “I built something 50 people actually use” is enormous in hiring
4. Consider adjacent roles — DevRel, solutions engineering, technical writing. These roles value junior technical skills AND human communication that AI can’t replace

But honestly? I feel like I’m giving them life vests while the ship is sinking. The structural problem @vp_eng_keisha described — fewer juniors now means fewer seniors later — isn’t something individual career advice can solve.

The Part Nobody Wants to Say Out Loud

We’re creating a two-tier system. Engineers who got in before 2023 have the experience, the networks, and the career trajectory. Engineers trying to get in after 2024 face a dramatically narrower funnel. And the divide will compound over time.

The 7 years of experience I have? They’re worth more now than they were two years ago, not because I’m better, but because the supply of people who got that same opportunity is shrinking. That should make me feel secure, but instead it makes me worried — because the seniors retiring in 2030 aren’t being replaced by the juniors we should be training today.

This isn’t just an industry problem. It’s a generational one.

[cto_michelle]

I’ve been reading this thread with the dual perspective of someone who has 25 years in this industry and who is currently scaling an engineering org from 50 to 120 engineers. Let me share the view from the C-suite, because the decisions being described here aren’t happening in a vacuum — they’re happening in boardrooms where I sit.

The Board Conversation Nobody Shares

Here’s what actually happens when junior hiring comes up in leadership discussions. The CFO shows a slide: “Senior engineer: $180K fully loaded, ships from day one. Junior engineer: $95K fully loaded, negative productivity for 6 months, break-even at 12 months, net positive at 18 months.” Then someone asks, “What about AI tools?” And suddenly the narrative shifts to, “Why invest in 18-month ramp-up when we can get 80% of that output from Copilot tomorrow?”

I’ve pushed back on this framing at my own company. But I want to be honest: the pushback is getting harder as AI capabilities improve quarter over quarter. The ROI case for junior hiring is a long-term argument in a market that rewards short-term efficiency.

Where I Agree — And Where I Don’t

@eng_director_luis, your medical residency model is the right analogy. Medicine recognized decades ago that you can’t just have attending physicians — you need a structured pipeline. But there’s a crucial difference: medical residencies are subsidized. Teaching hospitals receive federal GME funding specifically to offset the cost of training. Tech has no equivalent.

This is where I think our industry needs to get creative:

What if we treated junior hiring as R&D investment, not headcount cost?

At my company, we moved our junior engineering program budget from the “hiring” line item to “R&D and talent development.” That single accounting change transformed how leadership views the investment. It’s no longer competing against senior hires — it’s competing against conference budgets and training platforms. And in that context, investing $95K in a human who generates novel solutions, asks questions that challenge assumptions, and becomes a productive team member in 12 months looks like a bargain.

The Workforce Planning Math That Keeps Me Up at Night

I ran a model for my board last quarter. Key findings:

• Our average senior engineer tenure is 3.2 years
• We need to replace ~15 seniors annually through attrition
• External senior hiring takes 4.5 months average and costs $35K in recruiting fees
• Our internal junior-to-senior pipeline takes 3-4 years but costs 60% less per senior produced
• At current rates (zero junior hiring), we’ll face a 40% senior shortage by 2029

The “just hire seniors” strategy has a ceiling. There aren’t enough experienced engineers to go around — especially as every company simultaneously stopped training new ones. @alex_dev is right that this is a generational problem. From a strategic planning perspective, companies that maintain junior pipelines now will have a significant competitive advantage in 3-5 years when the talent crunch hits.

What I’m Actually Doing About It

I’m not going to pretend I have this fully solved. But here’s what we’ve implemented:

1. Structured apprenticeship program: 6-month rotations across three teams, with dedicated mentors. We hire 4 apprentices per cohort, twice per year. Not every apprentice converts, and that’s okay.

2. AI-augmented onboarding: New juniors get paired with AI tools AND human mentors. The AI handles the “how do I set up my environment” questions. The humans handle the “why did we architect it this way” questions. This has cut our ramp-up time from 6 months to 4.

3. Cross-functional junior roles: Some of our most successful junior hires aren’t pure developers. They’re in hybrid roles — developer + technical writer, developer + QA, developer + customer engineering. These roles are harder to automate away because they require human judgment at the intersection of disciplines.

4. Executive accountability: Junior hiring targets are part of my OKRs, not optional. If we don’t maintain pipeline diversity — in experience level, in background, in perspective — I’ve failed at a core part of my job.

The Uncomfortable Truth for CTOs

@vp_eng_keisha called out the hypocrisy, and she’s right. Every CTO in tech today benefited from someone taking a chance on them when they were junior. I was hired at Microsoft in 2001 as an entry-level SDE with a CS degree and not much else. Someone invested in my development. Someone was patient when I shipped bugs. That investment is worth millions in value I’ve created since.

If we collectively decide that investment is no longer worth making, we’re not just hollowing out the career ladder. We’re pulling up the drawbridge after we crossed it.

The companies that will win the next decade aren’t the ones with the most AI tools. They’re the ones with the deepest bench of experienced engineers who learned their craft through real-world practice. You can’t shortcut that with a language model.