Insurance Underwriting & Claims Expert – AI Evaluation
Leading AI Lab (via Braintrust) · United States, Canada, Austria, Belgium, Bulgaria, Croatia, Cyprus, Czechia, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Poland, Portugal, Romania, Singapore, Slovakia, Slovenia, Spain, Sweden, United Kingdom · Remote
- Pay
- $75–90/hr
- Commitment
- hourly
- Hours / week
- ~20
- Source
- braintrust
Is this job right for you?
Get job recommendations matched to your experience, first access to exclusive opportunities, and no ads.
Get jobs matched to your experienceAbout this role
We’re looking for experienced underwriters and claims professionals to create realistic cases for testing AI. You’ll draw on the kinds of files you’ve handled in practice, such as a commercial property application that calls for a rating adjustment or a claim where an endorsement changes the coverage decision.
What you’ll do
- Build realistic underwriting or claims files using a fictional insurer’s rules. Include the information an expert would need to make a decision.
- Work through each file and document the decision, the evidence behind it, and the rule or policy wording that applies.
- Create the corresponding output, such as an issued policy with the relevant endorsements, a coverage letter, or a settlement recommendation.
- Include details that matter to the outcome, so each case tests whether the AI applied the rules correctly and reached a defensible decision.
What you’ll bring
- Hands-on experience in either underwriting or claims. For underwriting, you’ve taken applications through to issued policies, including rating and selecting clauses or endorsements. For claims, you’ve personally made coverage, routing, and settlement decisions.
- Experience in one or more insurance lines, such as commercial property, motor fleet, health, or life.
- Ideally, five or more years of relevant experience. Ten or more years with continued hands-on work is a plus.
- Experience writing or reviewing AI evaluation cases or training data is a plus.
Contract details
Remote freelance contract, approximately 20 hours per week, with an initial term of three months.
Eligible applicant countries
This role accepts applicants from:
- United States
- Canada
- Austria
- Belgium
- Bulgaria
- Croatia
- Cyprus
- Czechia
- Denmark
- Estonia
- Finland
- France
- Germany
- Greece
- Hungary
- Ireland
- Italy
- Latvia
- Lithuania
- Luxembourg
- Malta
- Netherlands
- Poland
- Portugal
- Romania
- Singapore
- Slovakia
- Slovenia
- Spain
- Sweden
- United Kingdom
Skills & domains
- ai-training

