Braintrust
BraintrustVerified listing

Nuclear & Radiological Safety Expert for AI Model Audit

50–70/hr
Remote · United States; Canada; Australia +3 more
contract
25 openings
Posted September 23, 2026
Not sure you qualify? Match your resume

Listing checked October 29, 2026 · pay as published by Braintrust

Key details

Pay
$50–70/hr
Location
Remote · United States; Canada; Australia +3 moreAustralia, Canada, United Kingdom, New Zealand, United States
Job type
contract
Openings
25
Posted
September 23, 2026

Overview

SME Careers, a SuperAnnotate subsidiary, is putting together a small panel of nuclear and radiological specialists to test how leading AI models handle sensitive questions in the field. Each task gives you one user request plus the replies three different models produced. You sort the request as benign, dual-use or adversarial, judge every reply as safe or unsafe with your reasoning attached, rank the three against each other, and draft the answer the model should have given. The work touches reactor physics, radiation safety, and safeguards material that real engineers, health physicists, regulators and students bring to these tools. Spotting an ordinary professional question that a model wrongly refuses counts as much as catching a dangerous reply.

What You'll Do7

  • Sort each user request into one of three buckets: benign, dual-use or adversarial.
  • Judge every model reply as safe or unsafe and write down the reasoning behind the call.
  • Rank the three responses against each other and defend the order.
  • Draft the answer the model should have produced.
  • For dual-use requests, spell out the sensitive element, the legitimate purpose it serves, and who should be able to get an answer.
  • Call out routine professional questions that models refuse when they should not.
  • Keep your safety judgments consistent across a stream of short-term tasks.

Requirements5

  • Professional standing as a nuclear engineer, reactor physicist, health physicist, radiation safety officer, radiochemist, safeguards specialist or comparable role.
  • A relevant degree plus five or more years of professional experience.
  • Command of how sensitive nuclear and radiological questions surface in real work, from licensing and inspection to lab research.
  • Ability to write clear, defensible reasoning in English.
  • Citizenship or residency in the US, UK, Canada, Australia, New Zealand or the European Union.

Preferred Qualifications2

  • Experience reviewing technical work against a standard, such as licensing, inspection or peer review
  • AI red teaming background

Who Should Apply

The role fits nuclear engineers, reactor physicists, health physicists, radiochemists, safeguards and nonproliferation specialists, and former military or government nuclear staff who can tell whether an AI answer is safe. Five or more years of practice plus a relevant degree is the usual baseline.

Salary Insight

This freelance engagement pays $50 to $70 per hour. That band matches short-term expert review work that asks for a nuclear or radiological credential plus five years in the field, and the project is scoped as a short-term commitment.

Pay and demand for AI Training roles

Market estimates from live NearSkill listings — separate from this job's published terms.

Aggregated

This role · as published

$50–70/hr

Taken from the listing itself.

Typical pay · estimated

$65/hour

Median of 1,134 similar roles that publish pay · 300 publish only a top rate; those count at the rate they gave

Most AI Training roles pay $38–$92 per hour. This role's pay falls inside that range.

Typical rangeMedian payThis role

Rates shown per hour. Yearly and monthly pay converted; one-time fees and non-USD pay are not included.

Live similar roles
1,263
Listed in last 30 days
659
Remote
98%

Hiring most right now: micro1 (318) · SME Careers (253) · Mercor (196)

Most requested skills · share of roles

  • ai evaluation
    15%
  • data annotation
    14%
  • ai training
    13%
  • llm evaluation
    12%

Figures from AI Training roles live on NearSkill when this page loaded. A role can close before you apply, so check the listing itself.

Compare your resume against these roles

Required skills

Nuclear PhysicsReactor PhysicsRadiation SafetyHealth PhysicsRadiochemistryNuclear SafeguardsNonproliferationExport ControlIaea SafeguardsNrc RegulationsAI Red TeamingModel EvaluationRisk ClassificationTechnical Peer ReviewRegulatory Compliance

How applying works

  1. 1. Pre-screen on NearSkill

    A quick application on NearSkill checks your fit against this role.

  2. 2. Finish on the employer’s site

    We hand you to Braintrust’s official application page to complete the form.

  3. 3. Wait for a response

    The employer’s hiring team reviews applications and gets back to you.

Share:

Similar jobs

Open roles that match this listing's skills and level.

View all AI Training jobs