P&C Underwriter & AI Risk Specialist

Obsidian; New York (NY); On-site; USD 55,000 - 103,000; Part time

Job summary

Mercor is partnering with a leading AI lab to train frontier models on high-quality insurance reasoning data.

We're hiring Multiline Property & Casualty Underwriters to design realistic risk-selection and underwriting scenarios, evaluate model outputs against established underwriting standards, and help shape how the next generation of AI reasons about insurability, pricing, coverage terms, and portfolio fit.

Qualifications

• 2+ years of professional P&C underwriting experience
• Underwritten at least one major segment (personal lines, commercial property, general liability, etc.)
• Understand underwriting lifecycle from submission to renewal
• Explain how account characteristics, jurisdiction, peril, loss experience impact decisions
• Strong written reasoning and risk judgment
• Separate underwriting from claims handling and actuarial analysis

Responsibilities

• Design realistic underwriting scenarios involving new business, renewals, exposure analysis, loss history, inspections, classification, pricing, limits, deductibles, exclusions, and referrals
• Create work products such as underwriting assessments, quote rationales, information requests, declination explanations, renewal strategies, and risk-improvement requirements
• Write high-quality, experienced-underwriter level golden reference responses
• Grade AI-generated responses against rubrics for accuracy, consistency, risk judgment, and guideline adherence
• Identify unsupported assumptions or incorrect policy term usage and inappropriate acceptance/declination decisions
• Provide written feedback to the research team to improve model behavior
• Participate in onboarding office hours and calibration sessions

Job description

Mercor is partnering with a leading AI lab to train frontier models on high-quality insurance reasoning data.

We're hiring Multiline Property & Casualty Underwriters to design realistic risk-selection and underwriting scenarios, evaluate model outputs against established underwriting standards, and help shape how the next generation of AI reasons about insurability, pricing, coverage terms, and portfolio fit.