Senior Analytics Engineer

The Senior Analytics Engineer will be responsible for a range of responsibilities, including (but not limited to):

1. Carry derived data points through from development through QA into production, including the logic that defines them and the tests each one has to pass before they are consumed downstream.
2. Write the data methodology and logic customers ultimately see: episode triggers, provider attribution, cost standardization, and quality measure numerators, denominators, and exclusions.
3. Manage and build upon Clarify’s reference data library. This spans physician and facility affiliation mappings, codesets and their derivative mappings (ICD, HCPCS, NDC, DRG, etc. and their mappings to service lines/higher level categories), geographical info, etc.

1. Prove a number before anyone else has to, using reconciliation SQL, version-over-version comparisons that show what moved, and impact analysis run before a logic change ships.
2. Turn one-off QA into standing coverage: the data tests, the comparisons against a known-good baseline rather than the previous run, and the completeness thresholds that have to pass before data reaches a customer.
3. Run the data refresh and delivery cadence for customer environments once ingestion is complete, working test failures to root cause with engineering and handing off validated groupings to data science and product.

1. Scope feature requests from product, clinical informatics, and customer teams, write the specification and acceptance criteria, decide what you do yourself and what goes to an engineer, and follow it to release.
2. Settle methodology that crosses teams, including attribution logic, referral definitions, and provider reference data, where data quality, run time, and scope get traded against each other.
3. Hold AI-assisted tooling to the same standard as your own work. Clarify uses agentic tooling for data sourcing, specification drafting, and triage, and this role checks generated output against source data and sets the review it has to pass.

We are always looking for new team members who will add to our culture and have a strong passion for impact. In particular, the Senior Analytics Engineer will have:

1. Bachelor's degree in a quantitative, technical, or health-related field, or equivalent practical experience
2. Five or more years in data analysis, healthcare analytics, or a comparable data-heavy role
3. Advanced SQL. You query large datasets without help, build a metric from raw source records, define its grain, and defend every value in it
4. Experience in the AWS ecosystem, experience with industry-leading AI models, especially in their use for code development
5. Working knowledge of healthcare data and the context around it, such as claims, provider, or network data, and what the codes on a claim represent
6. Experience with data in a production setting, where tables refresh on a schedule, other teams depend on your output, and changes move through QA before release
7. Clear writing. Specifications, acceptance criteria, and decision records an engineer can act on and an auditor can follow six months later

1. A high bar and the work ethic to hold it. You do whatever it takes to get there — chasing down the missing answer, filling the gap no one owns, staying with it until it is right. You expect the same standard from the work around you.
2. Ownership mindset. You treat the components you own as yours, and you carry a problem to its outcome rather than handing it off halfway.
3. Comfort with ambiguity. The problems may be clear but not yet well defined. You make a reasonable call, state your assumptions, and keep moving.
4. A bias toward action. You are a doer. When something needs to be built, validated, or unblocked, you go do it, and you prototype your way to an answer.

1. Depth in a healthcare domain such as episode grouping, risk adjustment, quality measurement, or provider and network reference data, including CMS or NCQA measure specifications, clinical value sets, or standards such as HL7, FHIR, and X12
2. BI and dashboarding tools, and a record of explaining an analysis to an audience that does not work in the data
3. Recurring data quality checks you set up yourself rather than one-off validations
4. Python/Pyspark experience