EXPOSED
This is a catch-all bucket — photonics, optical, robotics, energy, validation, systems and reliability engineers — so the modal worker is a degreed non-PE engineer splitting time between simulation/analysis on a screen, test-and-measurement in a lab or plant, and specification and report writing. The screen half (parametric calcs, tolerance studies, requirements documents, test-report drafting, data reduction, code for instrumentation) is exactly where AI is already usable, while the bench-and-field half — building the test rig, diagnosing why the prototype fails, signing off that a design is safe to build — holds. Most of these engineers work under the industrial exemption, so there is no license forcing a human signature on their output.
Roughly flat across the period, with year-to-year wobble.
Median pay $99,040 → $122,930 -0.7% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
+2.1% 158,800 → 162,100 on the projections basis
Growing, and only partly exposed
The BLS expects +2.1% more of these jobs by 2034, and at 48/100 the work is only partly exposed — some tasks are automatable, the core of the job is not. Nothing here is in tension.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~9,300 openings a year on average, including replacing people who leave.
EngineerBioengineerMetrologistTest EngineerCarbon AnalystEnergy ModelerField EngineerLaser EngineerOcean EngineerRobot OperatorSolar DesignerSolar EngineerDesign EngineerDevice EngineerEnergy EngineerHydrodynamicistMobile EngineerOptics EngineerPatent EngineerSystem EngineerCoastal EngineerControl EngineerDisplay EngineerHeating Engineer
The BLS uses Engineers, All Other for work that doesn't fit any named occupation, so it covers roles that have little in common with each other. Two consequences worth knowing before you read anything below:
Holding it up: judgment & accountability . Weakest point: liability shield .
Mixed — a routine tier and a judgment tier An 11 reflects the split where tolerance stacks, FEA/optical model setup, requirements traceability matrices, and test-report drafting are already being handed to AI-assisted tooling, while aligning a laser bench, instrumenting a prototype, and working out why a robot cell drifts after 400 cycles still demand someone in the lab — if the work were purely simulation and documentation it would sit near 5, and if it were mostly field commissioning it would clear 15.
Some physical or field component A 9 puts most of these engineers in a lab, test cell, or plant floor rather than an uncontrolled site: you do handle hardware, fixtures, oscilloscopes, and environmental chambers, but the conditions are engineered ones you control, unlike a wind-turbine or pipeline engineer climbing structures in weather, who would score in the mid-teens.
Certification preferred, not legally required A 7 is the industrial exemption in practice — an EIT or a degree plus an internal design-review sign-off is what gates your work, and the PE stamp that would push this past 12 is only needed by the minority of you touching public works or offering services to the public.
Meaningful discretion A 13 sits at the top of real discretion: you decide whether a marginal qualification result is a real failure or a fixture artifact, whether to accept a supplier deviation, and when a design is ready for release — calls that get scrutinised in review boards and root-cause investigations, but that a chief engineer or safety authority formally owns rather than you.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (11/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (7/20) is whether the law requires a licensed human to sign. Trust premium (8/20) is whether buyers specifically pay for a person. Judgment and accountability (13/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 28 of this occupation's 48 points (58%).
Embodiment (9/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 65/100, still EXPOSED.
Functional-safety and cyber regimes that require a named, competent human as accountable signer: ISO 26262/26580 safety-case sign-off, IEC 61508 functional safety manager, EU Machinery Regulation 2023/1230 and the AI Act requiring identified natural persons for conformity assessment and technical files on high-risk machinery. If audit practice hardens into 'AI-generated analysis may not constitute the evidentiary basis without human-attested verification', the signature becomes non-delegable.
Genuine two-tier structure: if parametric calcs, tolerance stacks, requirement drafting and test-report generation are absorbed, the residual is failure diagnosis on real hardware, test-rig design, deciding which anomaly matters, and choosing what to measure. That residual tier is the harder half and its share of the day rises mechanically without any legal change.
Narrowing or repeal of the state industrial exemption so that in-house engineering work on products affecting public safety must be sealed by a licensed PE — NCEES has repeatedly floated exemption reform, and states like Texas and Ohio have live debates. A narrower version already in motion: FAA Organization Designation Authorization reform after the 737 MAX (Aircraft Certification, Safety, and Accountability Act of 2020) putting named engineering unit members personally on the hook for compliance findings.
Post-incident regimes that name individual engineers in the accountability chain — e.g. NTSB/CSB recommendations, DOE nuclear/grid interconnection reviews, or an insurer requiring a named engineer of record on a validation protocol before writing product-liability coverage. Also plausible: FDA design-control expectations naming a responsible engineer for verification of AI-assisted design outputs.
Shift of role weight toward bench and field work — commissioning, hardware-in-the-loop bring-up, root-cause teardown, environmental and EMC test campaigns — as screen tasks compress. No new institution required; the bucket already includes heavy lab-resident specialties (photonics alignment, robotics integration, reliability HALT/HASS).
The limit. Trust premium has no realistic route here: buyers of this work are employers and OEM customers who purchase certified outcomes and test data, not a named human's craft, so there is no visible channel through which anyone pays extra for human authorship. The liability levers are also the only large ones, and industrial-exemption reform has been proposed and defeated repeatedly for decades — the safety-case and conformity-assessment route is far more likely than PE licensure spreading. Even fully realized, the occupation is too heterogeneous for any single lever to lift the whole bucket; photonics and nuclear-adjacent engineers could gain most of this while validation and data-reduction roles gain none.
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 9,230 | $151,340 +23% |
| Los Angeles-Long Beach-Anaheim, CA | 7,530 | $125,060 +2% |
| Detroit-Warren-Dearborn, MI | 5,800 | $126,250 +3% |
| New York-Newark-Jersey City, NY-NJ | 5,780 | $135,300 +10% |
| San Jose-Sunnyvale-Santa Clara, CA | 4,970 | $169,650 +38% |
| San Francisco-Oakland-Fremont, CA | 4,850 | $159,800 +30% |
| San Diego-Chula Vista-Carlsbad, CA | 3,570 | $132,790 +8% |
| Houston-Pasadena-The Woodlands, TX | 3,490 | $130,270 +6% |
| San Jose-Sunnyvale-Santa Clara, CA | 4,970 | $169,650 +38% |
| Lexington Park, MD | 1,020 | $166,230 +35% |
| Albuquerque, NM | 1,110 | $163,630 +33% |
AP reports a South Korean startup is recording skilled workers' manual techniques to train AI models that control robots.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.