Artificial Intelligence Lead

Function: Insurance (Operations, Products & Technology)

Experience: 8–12 years (with 4+ years in applied AI/ML and 2+ years leading delivery)

Role summary

We are hiring an AI/ML Lead to own how AI is applied across insurance

workflows—underwriting support, claims, policy servicing, document-heavy operations, and

customer/agent experiences. You will sit between business and technology: translate

operational problems into AI solutions, set delivery standards, and ensure models and

platforms actually move cycle time, accuracy, leakage, and customer experience—not just

PoCs.

You will lead a small AI/ML squad, partner with product, operations, risk/compliance, and

engineering, and take solutions from problem framing through production and adoption.

What you will do

Business partnership

Work with claims, underwriting, operations, and product leaders to identify high-

value AI use cases (document intake, extraction, decision support, fraud/leakage

signals, servicing automation, knowledge assistants).

Convert messy insurance processes into clear problem statements, success metrics,

and phased roadmaps.

Set expectations with stakeholders on accuracy, exceptions, human-in-the-loop, and

what AI will not do.

Communicate progress, risk, and ROI in business language; bring technology

constraints back to the business early.

Solution ownership (not model-shopping)

Design end-to-end AI solutions: data → processing → intelligence →

APIs/workflows → ops feedback loop.

Choose the right approach for the problem (classical ML, NLP, computer vision,

GenAI, hybrid) without locking the org to a single technique.

Own quality: evaluation design, error analysis, sampling, thresholding, and

continuous improvement from production feedback.

Ensure solutions are production-grade: latency, cost, reliability, fallbacks, auditability,

and PII/sensitive-data handling.

Insurance domain depth

Understand policy, claims, FNOL, endorsements, KYC/onboarding, correspondence,

and operational SLAs well enough to challenge requirements.

Design for insurance document reality: multi-page PDFs, scans, mixed quality,

structured + unstructured content, exceptions, and regulatory language.

Align AI outputs to downstream systems (core, claims, CRM, workflow) and to

underwriting/claims decisioning—not standalone demos.

Leadership & operating model

Lead engineers and data scientists: hiring bar, coaching, code/design reviews, delivery

cadence.

Coordinate with platform, data, security, and application teams on architecture,

environments, and release.

Define the AI delivery playbook: discovery, MVP, evaluation, UAT with ops, go-live,

monitoring, and model/process change control.

Build trust with risk, legal, and compliance on explainability, documentation, and

responsible use of AI in regulated workflows.

Manage vendors and internal build-vs-buy where it affects speed, cost, or control.

What we look for

Leadership

Proven ability to lead AI/ML delivery with mixed seniority; you have been the person

business and engineering both trust.

Comfortable in ambiguity: you frame the problem, sequence the work, and protect the

team from thrash.

Strong facilitation: workshops with ops SMEs, alignment with tech leads, executive-

ready updates.

Ownership mindset: you stay with a solution after launch until it is used and stable.

AI/ML craft (applied, not academic)

Hands-on history shipping AI/ML into production (not only notebooks or vendor

configuration).

Broad applied range: supervised ML, NLP/document intelligence, vision where

relevant, and GenAI systems (including retrieval, grounding, and structured outputs).

Judgment on when GenAI is the right tool vs. when simpler ML or rules + ML is safer

and cheaper.

Experience with evaluation, drift, cost/latency trade-offs, and production failure

modes.

Enough engineering fluency to partner on APIs, data pipelines, cloud, and

integration—without needing to be the deepest specialist in every library.

Insurance / regulated operations

Direct experience applying AI/ML in insurance (P&C, health, life, or related—claims,

UW, servicing, or distribution).

Familiarity with operational KPIs: TAT, straight-through processing, leakage, first-

pass accuracy, exception rates, CSAT/NPS.

Respect for audit, consent, data minimization, and model risk in a regulated

environment.

Working style

Python-centric AI delivery background; comfortable with modern cloud and

production services.

Can write a crisp PRD-style problem brief and a technical design that engineers can

execute.

Bias to measurable outcomes over tool lists.

Nice to have

Built or scaled an AI CoE / chapter in a carrier, TPA, broker, or insurtech.

Experience with document-heavy operations (policy packs, medicals, invoices, IDs,

correspondence).

Exposure to MLOps/governance (experiment tracking, versioning, CI for models,

access control).

Prior coordination of offshore/onshore or vendor + in-house hybrid teams.

Success in 12 months

A prioritized insurance AI roadmap agreed with business and tech, with 2–3

production use cases live and measured.

Clear operating rhythm: intake, evaluation, production SLAs, and a feedback loop

with operations.

A team that can deliver without heroics; stakeholders who can explain what the AI

does, when it fails, and who owns the exception.

How this differs from an individual-contributor engineer role

This is a lead seat. Stack choices (frameworks, serving, OCR engines, vector stores, etc.) are

means, not the job. The job is: pick the right insurance problems, align business and

technology, ship reliable AI into workflows, and grow the team that keeps improving them.