EXPOSED
The modal biomedical engineer splits time between screen work that AI already does well — literature review, design-history-file documentation, 510(k)/technical-file drafting, risk analysis tables, test-data analysis, CAD iteration — and physical work that it cannot: benchtop prototyping, cadaver and animal lab testing, biocompatibility and sterilization validation, and sitting in an OR watching a surgeon fight your device. Design controls under FDA QSR require a named human to approve verification and validation, which is a real accountability hook even though PE licensure is rare in this field. The junior documentation-and-analysis tier is the most exposed; the design-owner and clinical-interface tier holds.
Dipped in 2020, then grew past where it started.
Median pay $91,410 → $109,370 -4.3% 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
+5.2% 22,200 → 23,300 on the projections basis
Growing, and only partly exposed
The BLS expects +5.2% more of these jobs by 2034, and at 47/100 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.
~1,300 openings a year on average, including replacing people who leave.
EngineerBioengineerField EngineerYield EngineerGenetic EngineerMedical EngineerProcess EngineerClinical EngineerDialysis EngineerResearch EngineerBiomedical EngineerBioprocess EngineerOrthopedic DesignerSustaining EngineerBiochemical EngineerBiomedical ScientistStack Yield EngineerBiomaterials EngineerFermentation EngineerBiomechanical EngineerBioinformatics EngineerField Clinical EngineerPharmaceutical EngineerImaging Service Engineer
Holding it up: judgment & accountability . Weakest point: trust premium .
Mixed — a routine tier and a judgment tier An 11 reflects that the tasks eating the most hours — DHF documentation, 510(k) predicate comparison tables, FMEA population, statistical analysis of bench and clinical test data, parametric CAD revisions — are exactly what generative tools now draft in a first pass, while designing a fixture to load a stent to 400 million cycles, debugging why a sensor drifts only in blood and not saline, and interpreting a failed sterilization validation still require someone in the lab; it sits above the mixed midpoint but not near 14 because most of what fills a workweek at a device company is documents and CAD, not benchwork.
Some physical or field component A 10 covers regular but scheduled physical presence — benchtop fatigue and burst testing, cadaver lab sessions, pilot-line builds, cleanroom entry, OR observation of a case using your implant — in environments that are still controlled and instrumented, unlike a field service engineer troubleshooting an installed MRI in a hospital basement, which is why this does not reach 13.
Certification preferred, not legally required A 7 is set by FDA 21 CFR 820.30, which requires a named individual to sign design review, verification, and validation records and makes that signature discoverable in a recall or warning letter, but there is no state licence to lose and no personal PE seal on most device work, so the exposure is employment and enforcement risk rather than the revocable credential that puts anesthesiologists and structural PEs above 11.
Meaningful discretion A 12 comes from calls that have no procedural answer and real consequence — whether a design change is significant enough to trigger a new 510(k), whether a single anomalous test unit is a manufacturing artifact or a failure mode, where to set an acceptance criterion when no standard covers your device — but those calls go through design review boards, regulatory affairs, and clinical advisers before they bind anyone, which keeps it below the 14+ band held by people who decide alone and in the moment.
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 (7/20) is whether buyers specifically pay for a person. Judgment and accountability (12/20) 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 26 of this occupation's 47 points (55%).
Embodiment (10/20) 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.
Nuclear Engineers EXPOSED
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 63/100, still EXPOSED.
Genuine two-tier structure: if literature review, DHF assembly, 510(k)/technical-file drafting, risk-table population and test-data analysis are absorbed by tooling, the residual job is bench prototyping, human-factors/usability validation under IEC 62366, surgeon-interface iteration, and failure-mode adjudication — a task mix with materially lower automatability. Watch for headcount concentrating in senior design-owner titles while junior documentation roles disappear
Notified-body and FDA expectations under EU MDR Annex XIV and FDA's real-world-performance push assigning a named design owner continuing responsibility for post-market clinical follow-up decisions — recall vs. field-safety-notice vs. label change calls, made on ambiguous complaint signal. This is a consequential-call role AI can inform but cannot own
FDA's QMSR transition (harmonizing 21 CFR 820 with ISO 13485, effective Feb 2026) plus the draft guidance on AI used in device design/production: if FDA requires a named, qualified individual to attest that AI-generated design outputs, V&V protocols, and risk analyses were independently reviewed — with that attestation discoverable in enforcement and product-liability litigation — the design-owner signature becomes personally consequential rather than clerical
Growth in point-of-care and patient-specific manufacturing (3D-printed implants, hospital-based device fabrication under FDA's discussion paper on point-of-care manufacturing) putting engineers physically in sterile-processing, print-shop and OR settings for per-case verification
State PE licensure actually being enforced for biomedical/clinical engineering practice (e.g., a state board ruling that hospital-facing device modification and OR-deployed custom implants require a licensed PE seal, as NCEES's biomedical PE exam nominally supports); currently near-zero uptake, and the exam was discontinued for low volume
The limit. Trust premium has no plausible route: buyers are hospital procurement, notified bodies and OEM program managers who purchase validated designs and regulatory clearance, not a named human's judgment. No patient or purchaser selects a device because a person rather than a model drew it. Combined realistic upside leaves this occupation exposed — the levers protect the senior design-owner tier while the junior documentation tier is what the headcount actually is.
| Boston-Cambridge-Newton, MA-NH | 4,040 | $128,550 +18% |
| Minneapolis-St. Paul-Bloomington, MN-WI | 1,070 | $127,890 +17% |
| New York-Newark-Jersey City, NY-NJ | 980 | $110,830 +1% |
| Los Angeles-Long Beach-Anaheim, CA | 910 | $128,240 +17% |
| Seattle-Tacoma-Bellevue, WA | 810 | — |
| Dallas-Fort Worth-Arlington, TX | 740 | $99,070 -9% |
| San Francisco-Oakland-Fremont, CA | 650 | $140,660 +29% |
| Philadelphia-Camden-Wilmington, PA-NJ-DE-MD | 580 | $125,760 +15% |
| Phoenix-Mesa-Chandler, AZ | 450 | $141,230 +29% |
| San Francisco-Oakland-Fremont, CA | 650 | $140,660 +29% |
| San Jose-Sunnyvale-Santa Clara, CA | 270 | $136,450 +25% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 47. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
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.