Last week we talked about the weird split screen between layoff headlines and overall hiring growth. This week there's a different split worth paying attention to, and it cuts right through the middle of most people's career plans: entry-level tech hiring is shrinking fast, while senior cloud infrastructure roles are in a genuine, hard-to-fill shortage.

Entry-level postings dropped from 8.1% to 7.4% of the total IT job mix year over year, while senior-level postings climbed from 38.8% to 43.1%. That's not a small wobble, that's a meaningful reshaping of where the jobs actually are. About two-thirds of enterprises say they're reducing entry-level hiring specifically because they're deploying AI, and over 90% report that junior roles have been changed or partially automated.

Think about what a junior engineer used to spend their first year doing: writing tests, fixing small bugs, moving data between APIs, learning the internal architecture by doing low-risk, well-defined work under supervision. A lot of that work is exactly what AI coding tools are good at now. That doesn't mean junior engineers are obsolete, but it does mean the traditional "learn by doing grunt work" on-ramp is narrower than it used to be.

At the same time, recruiters at hyperscalers are openly saying things like "we can't find enough qualified distributed systems engineers, cloud infra etc.," with one recruiter citing 20 open reqs sitting unfilled. Average pay for distributed systems engineers is now around $127,000, with a range of $98,000 to $157,000.

Why the shortage, when AI is supposedly automating engineering work? Because the things that make someone a strong distributed systems or cloud infra engineer are exactly the things AI still can't do well: system design tradeoffs, requirements clarification, production incident response under pressure, and judgment calls on security and compliance. That's earned experience, not a skill you can prompt your way into.

If you're early career, the old advice of "get any IT job and work your way up" still works, but the runway is shorter and the bar for standing out is higher. Companies that can't find senior people are still willing to hire and develop junior candidates who show the right signals, they're just pickier about which signals count. A cloud cert paired with a real, messy, self-built project (something you deployed, broke, and had to fix) says a lot more right now than a cert alone.

If you've got a couple years of cloud experience under your belt already, this is a genuinely good moment to point yourself toward distributed systems depth rather than staying broad. That means going past "I can deploy a VM and set up a VPC" and into the territory of multi-region architecture, failure modes, incident response, and the security and compliance reasoning behind your design choices. Those are the muscles that don't show up on a cert exam but absolutely show up in a senior interview.

Practical move: if you're mid-career, pick one production-style incident scenario (a region failover, a cascading service outage, a botched deploy) and write up how you'd design around it and respond to it, even hypothetically. That kind of artifact is exactly what's scarce right now, and it's a lot more persuasive in an interview than another bullet point that says "experienced with AWS."

The market isn't shrinking or growing uniformly. It's sorting people by how much judgment they can show, not just how many services they've touched.

Keep Learning and Building.