The talent market for senior engineers is broken in a way nobody is naming
The conversation is about AI, visas, and remote work. The actual story is a much older shift: the gap between 'engineer who can be hired' and 'engineer who can be trusted with a problem' has never been wider.
I have been hiring engineers, or helping to hire engineers, for the better part of a decade. The conversation about the talent market in that time has had a lot of characters — bootcamps, LeetCode, return-to-office, AI replacing juniors, AI not replacing juniors, junior-friendly job postings, “we can’t find senior engineers,” “we have too many senior engineers” — and through all of it, the thing I keep coming back to is a much less discussed observation.
The job market is not short on engineers. It is short on engineers who can be given a problem and trusted to come back with an answer.
Two markets, not one
The thing people mean when they say “the talent market” is, in practice, two different things. There is the interview market — the part that bootcamps, career changers, and most of the LinkedIn discourse is about — and there is the trust market, which is the part that actually decides what gets built and by whom. These markets are not very connected.
The interview market is, in objective terms, healthier than it has ever been. There are more competent junior engineers than at any point in history. The tools are better. The educational resources are better. The interview prep is, if anything, too good — a candidate with three months of focused study can pass a leetcode-medium screen at a tier-1 company in a way that was not possible in 2016. The pipeline is full.
The trust market is something else. It is, at most companies I work with, the binding constraint. The thing that limits what gets shipped is not the number of engineers with offers, it is the number of engineers whose judgement the leadership is willing to bet on. That number has not grown in step with the headcount.
Why the gap is widening
I think there are three reasons, and they are old reasons that have just gotten more acute.
The work has gotten more architectural. A senior engineer in 2014 was, in many companies, a strong individual contributor — the person who could ship the hardest single feature. A senior engineer in 2026 is, in the same companies, the person who can decide which features should be hard. The work has moved from “implement the right thing” to “decide what the right thing is,” and the second skill is much harder to assess, slower to develop, and more concentrated in a smaller number of people.
The cost of a wrong bet has gone up. When the system can be patched and redeployed in a day, you can afford to trust a less-experienced engineer with a bigger judgment call. When the system is load-bearing, regulated, or large enough that a wrong architectural choice compounds for years, the cost of the wrong call has gone up, and the willingness to make that call on a relatively new hire has gone down. This is rational. It is also, structurally, why “we can’t find senior engineers” is a complaint that exists in an industry that is, by headcount, in surplus.
The interview does not test for what gets trusted. I do not think this is a fixable problem. I think the thing that makes someone trustworthy on a real problem — the willingness to say “I don’t know,” the instinct to surface a risk before being asked, the ability to sit with ambiguity for a week — is, by its nature, the kind of thing that takes longer to assess than an interview loop is willing to spend. So we end up with proxies, and the proxies do not correlate very well with the thing.
What this means in practice
I want to be careful here, because this is the part of the post where the structural observation can curdle into advice, and the advice is usually for the wrong audience.
For the companies: if you are “short on senior talent,” the first question is not “how do we hire more seniors.” It is “what is it about our environment that makes it hard to develop seniors.” The answer is almost always the same: the seniors you have are spending all their time on the highest-leverage work, which is correct, and the mid-level engineers underneath them are not getting the reps. The fix is unromantic — slower growth, smaller teams, fewer parallel bets, more time per person. It is also the fix that works.
For the engineers: the gap is real, and it cuts both ways. If you are on the senior side, you have a lot more leverage than the headlines suggest, and a lot more options than “stay at your current company for the next decade.” If you are on the mid side, the strategy that has worked for the last ten years — get hired at a good company, get levelled up over four years — is, in 2026, harder to execute than it was in 2019, and the reason is mostly that fewer companies are running the apprenticeship model that the strategy depended on.
For everyone: the conversation about AI replacing engineers is, for now, mostly a story about the interview market, and the disruption to the trust market is going to be much slower and much weirder than the discourse suggests. The person you trust to decide what the right thing is, in 2030, is going to be a person who has, in addition to all the other things, learned when to delegate to a model and when to refuse to. That is a different skill than the one we have been hiring for. We don’t yet know what it looks like at scale.
The part I am uncertain about
I will close with the part I am least sure of. It is possible that the trust market is not actually broken, and what I am observing is the normal cyclical difficulty of finding senior talent, and we are just in a down-cycle the way the industry is always in a down-cycle. The 2010s felt like an open market. The 2020s feel closed. The 2030s will feel like something else.
But I don’t think the cyclical reading is quite right, because the kinds of jobs that are hardest to fill are the kinds that didn’t really exist in 2014. The person who can lead a refactor of a 10-year-old system, or who can be trusted to make a public API decision, or who can write the design doc that ten other engineers will then build from — that person was always rare, and they are rarer now, in absolute terms relative to the demand, than they have ever been.
I don’t have a tidy ending for this. I just think it is the conversation we should be having, and we are mostly not.