Freelance Clinicians (Benefits)

League Inc.

5d•Remote

About The Position

League is building the infrastructure behind our Benefits Navigation Agent — the AI that helps members understand and use their health benefits — and we're growing our panel of expert reviewers. The through-line is domain judgment: the situational, real-world reasoning that determines the guidance of our agent gives members is accurate, complete, and trustworthy. You'll help validate the rubrics we use to assess the agent's answers, grade responses against accuracy and coverage standards, and pressure-test the scenarios that keep benefits guidance correct and trustworthy. This is a great fit for benefits and insurance professionals who want flexible, remote work — and who want their expertise to shape a product while it's still being built.

Requirements

Meaningful hands-on experience in one or more of the specializations above (e.g., benefits administration, plan design, payer or claims operations, benefits regulatory/compliance, member advocacy, or supplemental/point-solution expertise)
Based in Canada or the United States, with working knowledge of the relevant benefits and insurance landscape
Strong command of where accurate benefits guidance ends and regulated advice (legal, tax, clinical) begins
Reliable, detail-oriented, and dependable on deadline-driven async work
Ability to access and operate within Google Workspace
Demonstrated experience using AI tools in a practical, responsible way
Curiosity and openness to experimenting with new technologies
Ability to balance efficiency with quality and sound judgment

Nice To Haves

Prior experience with evaluation or validation of AI products, or structured content/quality review
Familiarity with self-funded vs. fully-insured plan mechanics, accumulators, coordination of benefits, or claims adjudication
Genuine curiosity and interest about the role of AI in benefits navigation

Responsibilities

Scenario design and pressure-testing — building and stress-testing the member scenarios the agent is evaluated against, so the rubric reflects the questions members actually ask and the failure modes that matter most.
AI agent evaluation — supporting the development and validation of rubrics, grading AI-generated benefits guidance against structured accuracy and coverage standards, flagging where the agent misstates coverage or oversteps into advice it shouldn't give, and providing written feedback on how responses can be improved.
Subject-matter input — bringing a benefits, claims, or regulatory perspective to how we define correctness, coverage, and the hard-fail boundaries the agent must never cross.