Healthcare Data Analyst

AirPay•New York, NY

About The Position

AirPay automates dental insurance verification and revenue cycle workflows for dental practices. We process millions of eligibility and benefit transactions annually, sourcing data from payer portals, EDI pipelines, and direct connections. Our data team sits at the intersection of scraper engineering, payer operations, and product, turning raw benefit data into actionable outputs for practices and their patients. We’re looking for someone whose job is to find what’s silently wrong or missing in our benefit data, before a practice or a patient does. This isn’t a reporting role and it isn’t a dashboard role: the deliverable is catching the problem everyone else’s tidy summary missed. Coverage that reads “active” but lapsed last month. A field that comes back blank and gets quietly treated as “no benefit.” A payer whose data looks clean until you notice an entire location’s records are null. You’ll own a slice of our verification pipeline (a set of payers, a data domain, a monitoring surface) and run it: surface what actually matters, with evidence, without needing every anomaly pointed out to you first. You’ll report to the Head of Data and work as a peer-level collaborator with engineering, customer success, and operations.

Requirements

3 to 5 years where finding what’s missing or wrong was the job: data quality, completeness and validation, data audit, reconciliation, or investigative analytics. If your prior work was producing clean reports and dashboards, this is a different role. We’re hiring the person who finds what those reports leave out.
Healthcare data, payer analytics, insurance operations, or RCM background: Dental not required, but you need to understand how benefit structures work (coinsurance, deductibles, frequency limits, network tiers).
Strong data skills: comfortable in SQL and Excel/Sheets, can write a structured query, build a pivot, and sanity-check a dataset without hand-holding. More important than tooling breadth is the instinct to distrust a clean-looking result and dig until you know why it’s clean.
Self-directed ownership of an area. Point you at a payer, a domain, or a surface, and you run it: scope the questions, find the issues, frame them for a business audience, without per-step direction.
Clear written communication: Output is findings documents and structured reports, not just raw data.

Nice To Haves

EDI X12 familiarity (270/271 eligibility transactions especially).
Experience at a health tech company, payer, or benefits administrator.
Comfort with Python or scripting for data manipulation.
Experience with parser output, portal data, or unstructured benefit text.

Responsibilities

Run structured investigations into benefit data anomalies across 100+ payers. Triggers are both scheduled (cohort comparisons between data sources: portal-scraped vs EDI vs internal estimates) and signal-driven (customer escalations, operational reports). Output is engineer-actionable findings documents: what’s broken, where, what evidence, and critically, what’s wrong that nobody flagged. Common patterns include coverage-percentage inversions, blank-fill in returned data passed off as a real answer, frequency-normalization gaps (such as “12 months” vs “1 service year”), and stale data carrying across plan changes.
Own a monitoring surface and define what “healthy” looks like for it, then catch when reality drifts from it. The hard part isn’t building the dashboard; it’s noticing the gap the dashboard doesn’t show.
Own the weekly pattern analysis on customer-issue tickets. Categorize ticket types, track volume trends, identify recurring root causes, and surface signal that drives parser fixes, product changes, or QA process updates.
When a payer behaves unexpectedly or a hypothesis needs checking, you scope and run the investigation and write up the findings, starting from a vague “something looks off,” not a pre-scoped ticket.