Mathematical Science Occupations, All Other: EXPOSED

EXPOSED verdict contested

This is a residual category — model validators, quantitative analysts, cryptologic and biostatistical specialists who don't fit the statistician, actuary, or data-scientist codes — and the modal worker spends the day writing and testing code, cleaning data, fitting and documenting models, and producing technical memos. Those are exactly the artifacts LLMs now draft at usable quality, and nothing about the work is physical or licensed. What holds is the accountable end: choosing formulations under ambiguity, deciding whether a model is fit for a consequential decision, and signing the validation report someone in risk or regulatory will lean on.

10-year outlook: Routine quantitative production work in this residual category compresses hard over ten years, while a smaller, better-paid tier of model validators and methodology owners in regulated settings holds or grows.

contested two runs, 34 and 31 of 100 — unresolved

Score — 34/100 resistance

Five dimensions, 0–20 each, summed. Higher means more protected. The arithmetic is shown so you can check it: 9 + 3 + 3 + 7 + 12 = 34.

Task resistance 9/20

Mixed — a routine tier and a judgment tier. At 9 the split is visible inside a single day: the derivation setup, the choice of estimator for a badly-specified problem, and the argument for why a stress scenario is binding still take a person, but the Python that implements it, the backtest harness, the sensitivity tables and the SR 11-7 style write-up are now first-drafted by tooling faster than you can type them.

Embodiment 3/20

Fully desk- and screen-based. A 3 reflects that the only physical constraints are a workstation and, for the cryptologic and defence-adjacent slice of this code, a SCIF or air-gapped terminal you must physically sit at — that limits where the work happens, not what the hands do.

Liability shield 3/20

No licence, no signature requirement. There is no licence to strip: unlike the actuary who signs a Statement of Actuarial Opinion under ASOP, a model validator's report is absorbed by the CRO or model risk committee, and an SOA/CFA credential or clearance is a hiring filter rather than a legal barrier to anyone else doing the work.

Trust premium 7/20

Some relationship component. At 7 you have a real internal relationship — the model owners and the traders or epidemiologists who consume your output know whether your challenge is credible, and examiners recognise repeat authors — but the deliverable travels as a document under an institution's name, and clients rarely ask for you personally.

Judgment & accountability 12/20

Meaningful discretion. 12 is the effective-approve call: deciding a model is unfit for a capital decision or that a proxy variable invalidates a published estimate is a discretionary judgment with real consequences, but it lands inside a governance chain with committee sign-off, documented methodology standards, and an escalation path above you rather than terminating with your name.

The verdict on this page is contested. We scored this occupation twice, independently, and the two runs disagreed: 34/100 — EXPOSED and 31/100 — COOKED. Combining them doesn't settle it — the answer changes depending on how the halfway points are rounded, which is a property of our arithmetic and not of the job. Rather than pick the side that suits us, we've left the original verdict in place and flagged it. Read the dimension scores below and the reasoning attached to each; on this page they carry the information, and the one-word label does not.

Confidence: medium · reviewed 2026-08-11 · how scoring works

Tasks already automatable

writing analysis and simulation scripts
data cleaning and feature construction
drafting technical documentation and model write-ups
routine parameter fitting and sensitivity runs
literature and method surveys

What survives

formulating an ill-posed problem into tractable math
model validation and fitness-for-purpose sign-off
defending methodology to regulators and auditors
translating quantitative findings for non-technical decision-makers
novel method development where no benchmark exists

Active moats: judgment, trust

How to future-proof this job

Move into independent model validation/model-risk review where a named human attests that a model is fit for its stated use — that attestation is the durable deliverable.
Anchor in a regulated domain (SR 11-7 model risk, FDA biostatistics submissions, clinical trial design) where methodology must be defended in front of examiners, not just published internally.
Own the problem-formulation step with stakeholders — take briefs from business or clinical owners and specify the objective, constraints, and error tolerances before any code exists.
Take the AI-generated analysis pipeline as input and specialize in adversarial checking: identifiability, leakage, distribution shift, and where the model must not be used.

Escape hatches — adjacent fields with better verdicts

Computed from U.S. Dept. of Labor O*NET skill profiles: high overlap with what you already do, materially higher resistance score.

Medical Scientists, Except Epidemiologists EXPOSED · 49/100 · you already have ~70% of the skill profile

Skills to close: Science, Operation and Control, Operations Analysis, Writing

Natural Sciences Managers EXPOSED · 53/100 · you already have ~62% of the skill profile

Skills to close: Management of Material Resources, Management of Financial Resources, Management of Personnel Resources, Troubleshooting

Biological Science Teachers, Postsecondary EXPOSED · 53/100 · you already have ~58% of the skill profile

Skills to close: Learning Strategies, Instructing, Science, Speaking

Field report — do you do this job?

Has AI actually changed your work?

Self-reported and unverified — a sentiment signal, not a survey. One response per person per occupation; you can change your answer.

From people who do this job

Nobody has filed one yet. If you do this work, you know things the rubric can't see.