When I staffed the Senate Commerce Committee, I sat through dozens of hearings where scientists described the missions they wanted to fly. An infrared telescope that could see the first galaxies. A probe that could survive the radiation belts of Jupiter. A Mars lander that could drill into permafrost. The scientists spoke with genuine passion, and the senators nodded along. But the people I watched most carefully were the ones who never testified: the systems engineers sitting three rows back, already doing math on whether the thermal requirements were compatible with the mass budget, or whether the power draw from one instrument would starve another. Those were the translators. And they are becoming dangerously scarce.
What Systems Engineers Actually Do (and Why It’s Invisible)
The term “systems engineer” gets thrown around loosely in aerospace. It can mean the person managing interface requirements between a propulsion system and a spacecraft bus, or the person who reconciles a science team’s wish list with the physical constraints of a launch vehicle’s payload fairing. The common thread is translation: converting what scientists want to observe, measure, or explore into hardware that can actually survive the trip and produce the data.
This work is not glamorous. It doesn’t produce the dramatic renderings that make it into congressional slide decks or the front page of a NASA press release. A systems engineer’s best day often looks like preventing a catastrophic incompatibility between two subsystems that were each designed brilliantly in isolation but would fail together. Their worst day looks like discovering that incompatibility after metal has already been cut.
The craft requires a rare cognitive hybrid. You need enough understanding of thermal physics to talk to the thermal team, enough understanding of optics to talk to the instrument designers, enough understanding of structural loads to talk to the mechanical engineers, and enough understanding of budget cycles to talk to program managers. You don’t need to be the best in any of those fields. You need to be fluent enough in all of them to spot the conflicts nobody else can see because they’re each looking at their own piece. I’m not an engineer myself, but I’ve spent enough time around these people to recognize what makes them different. They think in systems, not components.
Consider the engineering lineage behind the James Webb Space Telescope’s infrared detectors. Those detectors needed to operate at cryogenic temperatures while mounted on a structure that had to unfold itself autonomously a million miles from Earth. The science requirement (seeing faint infrared signatures from the early universe) was clear. The engineering requirement (keeping detectors extraordinarily cold while dealing with thermal loads from the spacecraft itself) was a systems problem that took decades of iteration to solve. The people who bridged that gap, who figured out that the science requirement and the thermal requirement and the deployment requirement and the mass requirement could all coexist in one instrument, were systems engineers. Without them, JWST is a collection of independently excellent technologies that don’t work together.
The Workforce Crisis Is Structural, Not Cyclical
The aerospace industry is facing what multiple analyses have described as a workforce crisis, and the systems engineering discipline is being hit from several directions at once. The experienced generation, the people who worked Cassini, who worked the Hubble servicing missions, who worked the early Artemis architecture studies, are retiring faster than they can transfer their knowledge. The incoming generation, while talented, is being pulled toward software, AI, and computing-adjacent fields where the starting salaries are higher and the career visibility is greater. The 2026 aerospace industry forecast from Aerospace Manufacturing and Design identifies workforce challenges as one of the critical factors shaping the sector’s near future. The problem isn’t just raw numbers. It’s that the specific type of engineering mind needed for systems work, the integrative thinker, the translator, is the hardest to produce through a standard curriculum because the skill emerges from experience as much as from coursework. You can teach someone orbital mechanics in a semester. Teaching someone to feel the wrongness of a requirements document that’s internally inconsistent takes years of working on actual flight programs.
Some workforce shortages are cyclical. Demand spikes, supply catches up, equilibrium returns. The systems engineering gap is different because it’s rooted in how the aerospace sector is organized and how engineers build careers.
The training pipeline alone presents an enormous challenge. Industry veterans often suggest that a systems engineer typically needs 8 to 15 years of progressively complex program experience before they’re trusted to lead integration on a flagship mission. That’s not a training gap you close with a boot camp or a certificate program. It requires sustained investment in mentorship, in allowing mid-career engineers to work alongside senior ones on real hardware, and in keeping institutional continuity on programs that last long enough for that learning to happen. When programs get canceled, restructured, or stretched out over decades (as many NASA programs have been), the continuity breaks. Young engineers rotate off. Senior engineers retire without successors. Knowledge leaves the building.
Meanwhile, the competition for these minds is intensifying from multiple directions. Michigan Technological University’s fall 2025 data shows strong enrollment growth in engineering programs, with particularly robust interest in computing and interdisciplinary fields. Purdue’s Computes initiative represents a major expansion in computing and semiconductor research capacity. These are worthy priorities. But the student who might have become a systems engineer, the one who can hold five competing technical requirements in her head and find the design space where they all survive, is increasingly likely to become a machine learning researcher instead. Not because systems engineering is less important, but because the institutional incentives point elsewhere. A 25-year-old with strong analytical skills and a willingness to learn systems thinking can make significantly more in tech, finance, or defense software. The aerospace sector has never competed on salary with its talent competitors. It competed on mission: the chance to work on something that goes to space. That’s a powerful motivator, but it has limits, especially when student loan balances are real and Silicon Valley recruiters are aggressive.
The defense sector adds another structural drain. AeroVironment’s $200 million acquisition of ESAero, an electric propulsion and systems design firm, illustrates how defense companies are buying not just technology but engineering teams. When a small firm with deep systems integration expertise gets acquired by a defense prime, those engineers don’t disappear, but they’re no longer available for civil space programs. The defense budget is growing. Defense contractors pay well. And the security clearance requirements create a one-way valve: once an engineer is cleared and embedded in classified programs, moving back to civil space becomes logistically and financially unappealing.
As Space Daily reported in its coverage of the 2026 Space Symposium’s real agenda, the workforce gap was a recurring theme among industry leaders and government officials. The conversation keeps happening. The solutions keep not arriving at scale.
The Consequences Are Already Visible in Specific Programs
If this were an abstract problem, the kind of thing that shows up in workforce reports and gets discussed at panels before everyone goes back to their day jobs, I wouldn’t be writing about it. But the consequences are already showing up in program performance, and the pattern is unmistakable once you know where to look.
JWST itself is the most expensive example. The telescope’s science is spectacular, but the program’s cost grew from roughly $1 billion at its initial estimate to nearly $10 billion at launch. Multiple independent reviews identified systems engineering shortfalls as a core driver: interfaces between subsystems that weren’t fully reconciled early enough, integration complexities that weren’t staffed adequately during formulation, problems that compounded silently through years of design work before erupting during integration and test. The instrument teams were excellent. The component engineers were excellent. But the connective tissue, the systems work that binds brilliant subsystems into a functioning whole, was stretched too thin during the phases when catching problems is cheap.
NASA’s Mars Science Laboratory, the mission that delivered the Curiosity rover, tells a related story. The program overran its budget by roughly $900 million, and the independent review attributed significant cost growth to integration challenges: the sample handling system, the heat shield, and the landing system each met their individual requirements but created cascading interface problems when brought together. A more experienced, better-staffed systems engineering team in the early design phase might have caught those incompatibilities before they became billion-dollar fixes.
More recently, NASA’s own internal reviews of the Artemis campaign’s Space Launch System and Orion programs have consistently identified inadequate early-phase systems engineering oversight as a contributor to schedule slips and cost growth. The pattern is remarkably consistent across decades and programs: ambitious science goals, excellent instrument teams, excellent component engineers, but a systems engineering function that is understaffed, too junior, or brought in too late to catch the incompatibilities baked in during Phase A. By the time problems surface during integration and test, fixing them requires redesign, which requires money and time the budget didn’t account for.
The irony is that systems engineering is the cheapest insurance a program can buy. A well-staffed systems engineering team in the early design phase costs a fraction of what a redesign costs in the integration phase. But in Washington’s budget logic, the early-phase costs are easy to cut because the consequences are invisible at the time. I watched this dynamic play out repeatedly during my years on the Hill. When a program needed to reduce its Phase A budget to fit within a subcommittee’s allocation mark, systems engineering labor was often the first thing squeezed, because the science instruments are politically untouchable and the hardware contracts are spread across key congressional districts. Nobody lobbies for systems engineering hours.
What a Fix Would Actually Require
Acknowledging the problem is the easy part. Everyone in the aerospace community agrees that systems engineering capacity is eroding. The harder question is what to do about it, and here the conversation usually stalls because the real fixes require sustained institutional commitment, exactly the thing Washington is worst at providing.
The first requirement is protected investment in mentorship. That means keeping senior systems engineers on contract specifically to mentor mid-career engineers, even when the program they’re assigned to doesn’t have immediate integration work. This looks like waste in a cost-accounting framework. It’s actually insurance. NASA’s centers used to do this more effectively when the civil service workforce was larger and the institutional knowledge base was thicker. As more work has moved to contractors, the mentorship function has weakened because contractors are incentivized to bill hours against specific tasks, not to invest in developing the next generation of integrative thinkers for the broader industry.
The second requirement is curricular innovation. Universities are increasingly treating AI and computing as powerful instruments that demand both skill and responsibility, insisting students develop deep understanding using human creativity and judgment to unlock the full potential of the technology. This is a useful philosophical framework. But it needs to be paired with specific systems engineering coursework that teaches students to think across subsystem boundaries. The best programs do this through capstone projects where student teams have to integrate competing requirements into a single design, simulating the real constraints of a flight program. More universities need to offer this, and more aerospace employers need to signal that they value it in hiring.
The third requirement is compensation reform. The aerospace sector doesn’t need to match Google salaries. But it needs to close the gap enough that the mission appeal does the rest. Some of this is happening through retention bonuses at defense contractors and NASA centers, but it’s happening unevenly and often too late, after the most talented mid-career engineers have already left.
The fourth, and politically hardest, requirement is program stability. Systems engineers are made by working on programs that last long enough for them to see a full lifecycle, from requirements through design, integration, test, launch, and operations. When programs are perpetually threatened with cancellation, restructured every appropriations cycle, or stretched out so far that the design team turns over completely between Phase B and Phase C, the learning doesn’t happen. This is a problem I’ve written about in other contexts: the way Washington’s budget process undermines the long-term investments it claims to support. In my recent piece on the Space Symposium, I noted how the gap between stated policy ambitions and actual budget commitments creates real friction on the ground. The workforce dimension of that friction is the least visible but potentially the most damaging.
The Translators Can’t Wait
What concerns me most about the systems engineering shortage isn’t any individual program’s cost overrun or schedule slip. It’s the possibility that we’re losing something that’s genuinely hard to rebuild: a distributed institutional capability for turning scientific aspirations into real hardware that works in the unforgiving environment of space.
This capability doesn’t live in a document or a database. It lives in people, in the engineer who has seen three programs through integration and instinctively knows which interface is going to cause problems, in the team lead who recognizes from experience that two subsystem requirements are going to conflict even though both look reasonable in isolation. That kind of knowledge accumulates slowly and dissipates quickly. Once a critical mass of experienced systems engineers retires without passing their knowledge along, the institutional capability degrades in ways that are invisible until a program enters integration and starts discovering problems that a more experienced team would have caught years earlier.
The space community talks a lot about ambition. Artemis. Mars. Lunar Gateway. Commercial space stations. Exoplanet characterization. These are real programs with real budgets (however contested). But every one of them depends on the ability to take a scientific concept and translate it into hardware that survives launch loads, thermal extremes, radiation environments, and the thousand other failure modes that space imposes. That translation is done by people. Specific people with specific skills that take years to develop.
Global defense spending is booming, drawing experienced engineers into classified programs. AI and computing are absorbing the most analytically talented graduates. NASA’s civil service workforce is aging. Commercial space companies are growing fast but often lack the institutional depth to develop systems engineers internally. Each of these trends individually would be manageable. Together, they create a compounding problem.
The question isn’t whether we’ll notice. We will. The question is whether we’ll notice in time to do something about it, or whether we’ll notice the way we usually notice workforce problems in Washington: when a flagship mission blows through its cost cap and review boards point to systems engineering shortfalls as a contributing cause, and everyone nods gravely before turning their attention back to the politics of the next appropriations bill.
I’ve been in those rooms. The nodding is sincere. The follow-through is not. And the translators, the people who make the actual hardware possible, keep getting older, keep retiring, and keep not being replaced at the rate the mission manifest demands.
If we’re serious about the ambitions we keep funding, at least partially, we need to get serious about the people who turn those ambitions into flight-ready machines. That means protecting mentorship funding even when budgets tighten, reforming compensation before the talent is gone, stabilizing programs long enough for the next generation to learn, and treating systems engineering not as overhead to be squeezed but as the irreplaceable craft it is. The scientists will keep dreaming. The component engineers will keep innovating. But without the translators sitting three rows back, doing the quiet math that makes it all fit together, the dreams stay on the slide deck and the hardware stays on the ground. We still have time to change that. But the window is closing faster than Washington wants to admit, and unlike a launch window, this one doesn’t come back around.
Photo by Ron Lach on Pexels