Why Every Defense-Tech Careers Page Is Broken

We read ten of them, line by line — the open roles, the values lists, the perks, the things they beg for and the things they quietly leave out. A careers page is the one document a company can't fully spin: it has to be specific enough to make a stranger quit a good job, which means it leaks the real strategy. One of these pages told us a company is about to outgrow its plan. Most are broken in the same four ways. And the best of them are doing something the rest should steal.

The number that frames all of it: one of these companies added more than a thousand people in nine months and now runs north of 6,200. You cannot hire that fast on vibes. The page has to do real work.

The business behind the hardware. Not what they build — how they built it.

A change of format this issue. We've torn down three companies one at a time — Baykar, Helsing, Anduril. This time we tear down the artifact all three share, and the operator lessons hide in plain sight. Here are the six tells.

Every defense-tech careers page opens with a near-identical manifesto: mission, urgency, "build things that matter." The manifestos are interchangeable. The roles are not. The single most honest line on any careers page is the ratio of who they're hiring.

When a company that sold itself as "software-first" suddenly lists welders, manufacturing engineers, and production technicians next to its ML roles, the strategy has already changed — it has decided to own the metal, and the careers page admits it months before the press release does. The mix is the map. A wall of platform-engineering roles says "we're still building the core." A wall of deployed and field roles says "we're scaling delivery." A sudden block of manufacturing roles says "we're becoming a factory."

Takeaway: Audit your own careers page by role ratio, not by mission statement. The ratio is your strategy made literal — and if it doesn't match the strategy you say you have, one of the two is lying.

Helsing's careers page leads with a sentence most companies would never risk: it wants "people with their heart in the right place," who "share our conviction that democratic values are worth protecting." Anduril's frames the job as transforming military capability and warns, in effect, that you'll move fast and own your work. Read cynically, that's filler. Read correctly, it's a filter.

Mission-explicit pages are doing deliberate screening: they put the most polarizing thing about the work at the very top so that ambivalent candidates self-select out before they ever apply. In a sector where a wrong-fit hire churns in months, a page that repels the wrong people is doing more work than a page that attracts everyone. The repulsion is the product.

Takeaway: Put the thing that would scare off a bad-fit candidate above the perks, not below them. A careers page that doesn't repel anyone isn't filtering anyone.

Look for one specific role: the deployed, field, or customer-facing engineer. Palantir invented the "forward deployed engineer," and the pattern has spread — Anduril and the big AI labs now hire the role by name, because software sold into hard, security-sensitive environments needs deep customization after the sale, not before it.

Its presence or absence on a careers page is a strategy X-ray. A company selling complex systems into demanding customers with zero deployed-engineer roles doesn't yet understand its own delivery model — it's still pretending the product ships itself. A company with a growing block of them has admitted the truth: the last mile is a headcount problem, and it's staffing for it.

Takeaway: Staff the seam between product and customer before you're forced to. The deployed-engineer role isn't a title; it's a decision about who eats the last mile.

Here's the first of the four ways these pages break. A careers page that leads with free lunch, unlimited PTO, and a wellness stipend is advertising to everyone, which means it's advertising to no one. Commodity perks are noise; every company has them, so they carry zero information and select for candidates optimizing on comfort.

The pages that work sell the work instead. Helsing offers "the freedom to solve complex challenges." Anduril sells ownership, pace, and "systems that matter." Those aren't perks; they're scope. The best engineers in this sector are not bought with snacks — they're bought with the size of the problem and the amount of rope. A page that buries scope under benefits is fishing in the wrong pond.

Takeaway: Sell the problem, not the ping-pong table. Lead with scope and ownership; the people who pick you for the perks are the people who'll leave for better ones.

The second break: what a careers page leaves out is louder than what it prints. No manufacturing roles means you don't make anything yet. A single office means you're a single-market company whatever the mission copy claims. And the requirement everyone buries — citizenship, residency, or clearance gates — hidden three clicks deep instead of stated up front, quietly wastes the funnel and signals a company that hasn't decided how honest it wants to be.

Read the gaps and you read the constraints. A pan-European page with offices in four cities, like Helsing's, is telling you the company's real moat is talent access across borders. A page with one location and a sea of remote-software roles is telling you it hasn't bet on the physical world yet. The holes are the roadmap.

Takeaway: Audit your page for what's missing, then ask whether each gap is a choice or a blind spot. State the hard requirements on line one — burying them just means paying for the same rejection twice.

The third and fourth breaks live in the same place: the values section. Every page has one, and most are interchangeable posters — "integrity, ownership, impact" — that signal a copied template and prove nothing. A generic values list is negative information; it tells a sharp candidate you didn't think hard enough to be specific.

The exception proves the rule. Helsing's page says outright that its values "aren't posters on a wall; they're field-tested principles," then names them in operator language — "from first principles," "the truth is in the field," "whatever it takes." That only works because each one implies a concrete way the company behaves. The test is brutal and simple: if you can't attach a real operating example to a value, it's decoration, and decoration on a careers page reads as a tell that the culture is thinner than the copy.

Takeaway: Delete any value you can't pair with a specific story of the company living it. A short list with proof beats a long list with posters every time.

Issue 05. We've spent three issues on companies that tried to own everything — and a fourth on how they staff that ambition. So here's the inversion: what about the company that wins by refusing to do most of it? Next issue is a focused European builder that picked one narrow slice of the market, went deep instead of wide, and is quietly compounding while the giants sprawl. Its careers page is a fraction the size of Anduril's — and that's the whole strategy. We'll name it next issue: the case for doing less, on purpose.

Commander is the business behind the hardware. Not what they build — how they built it.

New transmission every two weeks. If it earned its place in your inbox, forward it to one operator who’s building something — that’s how Commander spreads.

COMMANDER · ISSUE 04 · commander.media · Istanbul