You will lead the engineering of production‑grade AI/GenAI services and agentic automations that uplift risk assessment, claims adjudication, and customer communications on a secure, multi‑tenant foundation. Key Outcomes & Engineering Responsibilities ● Own a 12–24‑month Insurance AI engineering roadmap across underwriting (L&A, P&C, Specialty), policy servicing, and claims—prioritized by business outcomes (STP, leakage reduction, TAT, loss ratio). ● Architect multi‑tenant AI services (LLM/RAG, risk/price models, adjudication engines) with API‑first interfaces, strong tenancy isolation, observability, and cost/latency SLOs; enable consumption by Insurance WorkDesk/agent portals and core ecosystems. ● Underwriting intelligence: ship services for submission ingestion, triage/prioritization, risk scoring, quote‑acceptance prediction, and document summarization; integrate with rules engines and rating/policy admin systems. ● Claims AI: deliver FNOL intake automation, AI triage, fraud/risk detection, and explainable adjudication for Life, P&C, and Health; standardize salvage/subrogation sub‑processes and omnichannel customer updates. ● Policy servicing & booking/binding: build GenAI‑assisted clause/wording libraries, template governance, and contract generation with maker‑checker workflows and full audit trails. ● Agentic insurance journeys: operationalize multi‑agent frameworks (e.g., Underwriting Assistant, Claim Adjudication) with guardrails (input/output filters, grounding, policy catalogs) and human‑in‑the‑loop controls. ● Document intelligence (IDP): embed classification, extraction, and redaction for applications, medicals, bills, loss evidence, and endorsements to reduce manual effort and errors. ● Customer communications: expose AI services that personalize and govern omnichannel communications (renewals, endorsements, claim letters) with template control and auto‑archival. ● Ecosystem integrations: design adapters for core platforms (e.g., Guidewire, Duck Creek), CRM, and data‑partner APIs; package deployables for marketplace motions where applicable. ● Establish insurance‑grade MLOps/LLMOps: model/data registries, offline/online evaluations (grounding, fairness, leakage impact), CI/CD, blue‑green/canary rollouts, rollback, run‑books; incident/SLA management. ● Build, coach, and scale a high‑performing team (applied science, ML/platform, evaluation & safety); drive design rigor, reliability, and measurable production impact. ### Requirements ● 10–12 years total; 5+ years leading AI/ML engineering teams shipping production AI in insurance (L&A/P&C/Specialty) across underwriting, policy servicing, or claims. ● Systems design depth: multi‑tenant AI services, vector/feature stores, streaming ETL, event architectures; observability and cost/performance optimization at scale. ● LLMs & decisioning: prompting, fine‑tuning, RAG; explainable decisioning aligned to underwriting/claims policies; propensity, fraud, and price‑sensitivity models. ● Document AI & IDP: OCR/ICR + layout models for medical records, bills, proofs; privacy/PII redaction; evidence packaging for audits. ● Domain fluency across Life & Annuity (policy issuance/underwriting, claims), P&C (policy booking/binding, claims), and Health (auto‑adjudication, pre‑auth). ● Ecosystem experience with insurance cores/platforms (e.g., Guidewire, Duck Creek), CRM, and data providers. ● Stakeholder leadership and communication; ability to explain model/platform trade‑offs to executives, regulators, and customers.