Data Engineer

Function Health•Austin, TX

16d

About The Position

Function Health is the AI operating system for health, designed to empower people to live 100 healthy years. We are redefining how individuals understand, measure, and improve their health by moving beyond the limitations of traditional care and enabling comprehensive, continuous insight into human biology. Function has been recognized as one of Fast Company’s Most Innovative Companies of 2024, and is venture-backed by Redpoint and other leading venture firms. Over half a million members have joined Function to take control of their health. Through comprehensive lab testing, MRI and CT imaging, longitudinal data, and guidance from AI and clinical teams, Function gives members a complete and continuous view of their health — and the clarity to act on it. Through acquisitions across imaging, mobile phlebotomy, and supplement tracking and optimization, Function is building an integrated health platform that spans the full path from testing to health clarity to daily action. Function recently announced a $298M Series B and is entering its next chapter of growth. As we scale, the quality and durability of our People systems, data, and insights will directly shape our ability to attract, retain, and support exceptional talent. We are growing our team and seeking out world-class talent that deeply believes in our mission to positively impact global health, has a relentless bias toward action, and a growth mindset. Function fosters a collaborative and dynamic environment where every day we build the future. At Function, the person who best understands a piece of data is rarely the person who can safely change it. A data scientist knows exactly how a biomarker should be derived. A clinician knows which reference range is wrong. Today, both have to file a ticket and wait for an engineer. We think that's a platform failure, not a process problem. Our job is to build the paved road that lets the subject matter expert make the change themselves – and lets an agent make it too – without anyone lying awake wondering about the validity or stability of what just shipped. That means the guardrails have to be real: declarative contracts instead of hand-rolled code, CI that diffs the data and not just the diff, validations that gate promotion, lineage that tells you who's downstream, and a rollback that takes one command. Get that right and both your humans and your agents get faster at the same time, for the same reason. That's the work. It's platform engineering, and the users are engineers.

Requirements

Built internal platform or infrastructure that other engineers actually adopted. You've felt the difference between shipping a tool and getting it used.
Operated production data or ML systems, on call for them, and fixed them under pressure.
Strong Python and SQL.
Comfortable in a lakehouse – we use Databricks; Snowflake or BigQuery translates fine.
Designed interfaces and schemas that other teams depend on, then evolved them without breaking those teams.
Thought hard about testing and CI for data or ML, where correctness is statistical and the failure is often silent.
1-4 years of engineering experience gets you here, but we care about what you've built, not the number.

Nice To Haves

Declarative pipeline frameworks (dbt, DLT, Dagster, Airflow)
Streaming (Kafka, Spark Structured Streaming)
Data contracts, data diffing, or lineage tooling
Terraform
Feature stores
MLOps and eval tooling
Agentic coding workflows
Healthcare PHI, or HIPAA experience

Responsibilities

Tracking infrastructure. The event pipeline behind product analytics, experimentation, and feature gates. Schemas that are enforced at the source, so a bad event never becomes a bad metric.
Data processing infrastructure. The Bronze → Silver → Gold layer in Databricks. Automated schema evolution, contract tests, backfills that aren't scary, freshness and volume monitors generated from the contract rather than bolted on after.
ML infrastructure. Feature computation and serving, training and eval pipelines, and the plumbing that gets model output back into the product — with the same testing and observability bar as everything else.
Cutting across all three. The self-service story. Templates, local dev and preview environments, policy-as-code for PHI, ownership routing for alerts, and progressive gates so an exploratory model ships freely while a member-facing one earns more scrutiny.