Role Level: This is a hands on individual-contributor role. Not suited for entry level or new graduate candidates.
Focus: The primary work is backend — a Python and SQL data platform running on Databricks, BigQuery, and PostgreSQL. Frontend and browser-extension work exists and you will touch it, but it is secondary. If you want to spend most of your time in React, this is not the role.
About the system. At the center of this role is a Python and SQL platform that generates and manages patient care-gap work for accountable care organizations. It reads CMS assignment files, Part A and Part B claims, Medicare eligibility data, and live hospital ADT and FHIR feeds out of BigQuery; it runs as scheduled Databricks jobs; and it writes to a PostgreSQL (Neon) database that a Next.js application and a Chrome extension read from.
Correctness matters more than velocity here. Records are append-only and fully audited, work a human has already done on a record must never be overwritten by an automated run, and a wrong join can change which care thousands of real patients get contacted about. Much of the engineering judgment in this job is about blast radius, idempotency, and being able to prove what a script did.
What You'll Do
- Own the opportunity-generation pipeline. Extend and operate the Python pipeline that turns Medicare claims, eligibility data, and hospital ADT feeds into prioritized, workable care-gap items for a portfolio of ACOs and commercial payors.
- Write heavy SQL against BigQuery. Query and tune multi-million-row datasets — assignment files, Part A and B claims and line items, eligibility responses, FHIR encounter and ADT feeds — including the deduplication and late-arriving-data handling those feeds require.
- Build and run Databricks jobs. Package pipeline scripts as scheduled jobs (Python script tasks, cluster and dependency configuration, DBFS, secret resolution), and run controlled one-off jobs against production when remediation is needed.
- Design and migrate PostgreSQL. Own schema, indexing, and query performance on Postgres (Neon): versioned SQL migrations, staging and historical mirror tables, JSONB metadata columns, read-replica usage for batch reads, and database-enforced immutability.
- Make generators idempotent and safe to re-run. Deterministic record IDs, upsert paths that preserve human work, supersede and carry-forward logic across reporting periods, status lifecycles, and an append-only audit log behind every state change.
- Protect production data. Build dry-run-then-commit scripts, volume abort thresholds, environment guards, backups, and rollback artifacts — and when something does go wrong, write the remediation that repairs the affected rows without destroying history.
- Extend automated data-quality monitoring. Add invariant and drift checks that run against live production data, wire up alerting, and investigate anomalies through to a written root cause.
- Test pipeline logic. pytest suites with ephemeral PostgreSQL fixtures and frozen clocks, plus the pre-commit gates that keep regressions out of a nightly run.
- Publish and defend data contracts. Document what each record type contains and what every field can hold, so the application team can build against it, and coordinate breaking changes across repos.
- Cross into the application when the work calls for it. Next.js, React, and TypeScript surfaces, REST APIs with role-based access control and HIPAA-compliant audit logging, Firebase Cloud Functions and other serverless services, and the Manifest V3 Chrome extension that integrates directly with web-based EHR systems.
- Debug end to end. Trace a wrong number on an application screen back through PostgreSQL, the generator that wrote it, the Databricks run it came from, and the upstream BigQuery feed.
- Collaborate across a small team and two repos. Participate in code review, work with the application and clinical-operations side on what the data needs to support, and help keep a large Python and TypeScript codebase maintainable.
- Handle Protected Health Information responsibly, with attention to performance, reliability, scalability, and security at every layer.
Required Skills
- 4+ years of professional software development experience. This is not an entry level or new graduate position.
- Strong Python for production data work, including pandas — scheduled batch jobs and command-line tools, not notebook-only experience.
- Advanced SQL: CTEs, window functions, large multi-table joins, aggregation correctness, and query optimization on large datasets.
- Strong PostgreSQL: schema design, indexing, EXPLAIN-driven tuning, transactions, upserts, and JSONB.
- Experience writing and applying SQL migrations against a live production database.
- Experience with a cloud data warehouse — BigQuery strongly preferred — including partitioning and an awareness of query cost and performance.
- Experience building and operating scheduled data pipelines: idempotent re-runs, backfills, duplicate and late-arriving source data, and reconciliation between systems.
- Experience with Databricks or Spark — jobs, clusters, dependency management, secrets, DBFS — or a comparable managed job platform you can map onto it.
- Demonstrated operational judgment with production data: dry runs, blast-radius estimates, auditability, and a rollback plan before you write.
- Experience debugging complex issues across multiple systems — warehouse, batch job, database, and application.
- Working ability in TypeScript with React and Next.js: enough to ship a frontend or API change yourself when the work requires it.
- Experience designing and building REST APIs.
- Comfort with a Git and pull-request workflow, including giving and receiving code review.
- Clear written communication — you can explain a data defect, its cause, and its fix in a document another engineer can act on.
- Strong problem solving skills and attention to detail.
Nice to Have
- Healthcare data experience: CMS Part A and Part B claims, Medicare beneficiary identifiers (MBI), HL7 v2 ADT feeds, FHIR, or eligibility services such as HETS.
- Value-based care domain knowledge: MSSP or ACO programs, risk adjustment and HCC coding, Annual Wellness Visits, transitional care management, or quality measures.
- Experience with healthcare technology, including EHR integrations (eClinicalWorks, Athena, Epic, or similar).
- Experience with Firebase Cloud Functions, Firestore, and Google Cloud Platform (IAM, service accounts, Secret Manager).
- Experience with Prisma ORM.
- Experience developing Chrome Extensions using Manifest V3.
- Data quality and observability work: anomaly detection, invariant checks, drift monitoring, alerting.
- Experience integrating LLMs such as Claude, OpenAI, or Gemini.
- Experience building speech to text or AI powered workflows.
- Experience building internal operator tooling — small consoles or TUIs for running, configuring, and monitoring pipeline jobs.
- Familiarity with modern Python tooling (uv, pytest plugins, pre-commit hook managers) and multi-package repositories.
- Familiarity with Rust for performance sensitive components, browser tooling, or systems programming.
- Familiarity with HIPAA compliance and secure handling of healthcare data.
Benefits:
- 401(k) 6% Match
- Bereavement leave
- Dental insurance
- Family leave
- Flexible spending account
- Health insurance
- Health savings account
- Life insurance
- Paid parental leave
- Parental leave
- Vision insurance
Pay: From $70,000.00 per year
Benefits:
- 401(k) matching
- Dental insurance
- Employee discount
- Flexible spending account
- Health insurance
- Health savings account
- Life insurance
- Paid time off
- Parental leave
- Professional development assistance
- Referral program
- Tuition reimbursement
- Vision insurance
Application Question(s):
- Have you worked on software handling PHI or in a HIPAA-regulated environment?
- Describe a production data pipeline you owned end to end: the data sources, the schedule, the volume, and what you did the first time it produced bad data.
- • What is your hands-on experience with BigQuery or another cloud data warehouse, and with Databricks or Spark?
- How comfortable are you writing and applying SQL migrations against a live production PostgreSQL database, and what precautions do you take?
Work Location: Remote
