Principal Data Lead — Provider Intelligence

J2 Health•New York, NY

1d•$200,000 - $240,000•Hybrid

About The Position

We're hiring a Principal-level data leader to own and build J2's provider-data "recommendation engine" — the part of our product that recommends contracting leads to health insurers, and distills a single source-of-truth provider dataset out of noisy, conflicting, multi-source data. This is a foundational role for the function. You'll set the vision for how J2 turns messy provider data into trustworthy, customer-facing data products; you'll be hands-on building it from day one; and you'll grow and lead the team behind it as it scales. The center of the job is judgment: looking at conflicting data and reasoning, fast and iteratively, toward an answer that makes sense to a real person. There's significant room to expand the surface area from here, for example, validating a customer's own provider data against J2's source of truth and flagging the records most likely to be wrong. This is a high-agency seat. We'll give you the problem and the leverage; you'll decide how to solve it and ship.

Requirements

A strong, opinionated point of view on how to apply LLMs to ambiguous data-judgment problems
Deep familiarity with the realities of messy, multi-source data and what it takes to reconcile it into something trustworthy
Enough business and product context to connect data work to customer value
Strong Python
dbt
Postgres
LLM frameworks
Look at conflicting data and reason quickly toward an output that makes sense to a human: speed, taste, and product sense are the qualities we're hiring for
Productize your own work: turn a data judgment into a shipped, customer-facing product, not an internal dashboard or a research artifact
Build hands-on today (strong Python; comfort across a modern data stack) while architecting for a team to scale behind you
Direct AI coding tools effectively so the leverage shows up in how much you ship
Product-minded, not research-minded. You're energized by "how do I make sense of this messy data and ship something useful," not by abstract modeling for its own sake.
Hungry and high-agency. This is the defining trait. You take an open-ended problem and run.
Founder energy. You want to do the work now and build the function — both, not one.
Commercially oriented. You instinctively ask whether the result actually makes sense and moves the customer, not just whether the job succeeded.
Senior / principal level, ~5–7 years is the sweet spot.
Real startup, 0-to-1 and 1-to-n product experience. You've built something from nothing and helped it grow.
A track record of shipping data products that customers used, not internal BI or published research.

Responsibilities

The provider-data recommendation engine end to end, from the conceptual model of "what is true about this provider" through to the recommendations customers act on
Distilling a source-of-truth provider dataset from many noisy, disagreeing sources, and being accountable for whether the output is right and useful, not just whether the pipeline runs
Bringing LLMs to bear on ambiguous data-judgment problems, directing AI coding tools to do the building so you and your PM partner can spend your energy on the hard judgment calls
Shipping iteratively and putting data products in front of customers, then sharpening them based on what actually drives value
Building and leading the function over time, hiring behind you and scaling a team as the work grows