Senior Insurance AI Benchmark Lead (Hybrid Bay Area)

Wintermeyer Ventures; San Francisco (CA); Hybride; USD 83 000 - 138 000; Plein temps

Résumé du poste

Wintermeyer Ventures is seeking a senior AI research specialist to join a generative AI team focused on improving how models reason about insurance and actuarial work. You will bring professional domain judgment to task design, model evaluation, and benchmark development.

Role involves writing task specifications, creating reference solutions, and designing domain-specific challenges while collaborating with researchers to turn expert judgment into clear evaluation criteria.

Qualifications

• 4+ years of substantive experience in insurance or related field.
• Actuarial credential (FSA/ASA/FCAS/ACAS) or active state insurance license.
• Senior-level experience with ownership of an insurance book, reserve, or claims portfolio.
• Specialization in life, group, property & casualty, health, reinsurance, pricing & reserving, ERM, or underwriting.
• Experience using large language models and strong judgment in insurance reasoning.
• Experience writing task specifications, reference solutions, or evaluation criteria.

Responsabilités

• Review insurance tasks and model outputs for missing behaviors and flawed reasoning.
• Write instruction specifications and create reference solutions for insurance problems.
• Design challenging domain-specific tasks and evaluation benchmarks.
• Collaborate with researchers to translate expert judgment into clear criteria.

Connaissances

Insurance domain knowledge

LLM experience

Evaluation criteria design

Task specifications writing

Formation

Actuarial science / mathematics / statistics degree

Description du poste

Wintermeyer Ventures is seeking a senior AI research specialist to join a generative AI team focused on improving how models reason about insurance and actuarial work. You will bring professional domain judgment to task design, model evaluation, and benchmark development.

Role involves writing task specifications, creating reference solutions, and designing domain-specific challenges while collaborating with researchers to turn expert judgment into clear evaluation criteria.