Analytics Engineer, Data Platform

Upside•Washington, DC

1d•$149,000 - $180,000•Hybrid

About The Position

Five million people use Upside to earn cash back on gas, groceries, and dining. The offers they see, the lifecycle messages they get, and the partner launches behind both all run on data models. You'll own a set of those. Not just building them. Deciding how they should be shaped, testing them, monitoring them, and documenting them well enough that someone else can safely build on your work. You'll sit close to Marketing, Data, and MarTech, so a lot of the job is turning a messy question into something concrete and trustworthy. Team of six. Snowflake, dbt, Dagster, AWS.

Requirements

Around 3–5 years in data or analytics engineering, or comparable work under a different title
Fluent in SQL, comfortable with window functions and complex joins, and think about query performance without being asked
Owned dbt models in a version-controlled repo; conventions, tests, CI, and the occasional cleanup of someone else's tangle
Know Python well enough to work in orchestration, transformations, and tests
Have an opinion on modeling tradeoffs (dimensional vs. one big table) and can explain which you'd pick and why
Can explain a technical decision to a marketer and an engineer in the same meeting, and adjust how you say it for each
Worked in Snowflake, or a comparable warehouse you could translate from

Nice To Haves

Marketing, growth, or lifecycle data: events, attribution, experimentation, or tools like Braze, Iterable, or Segment
Dagster, Airflow, or another modern orchestrator
CI/CD for data, data governance, or cost-conscious warehouse design
Supporting ML workflows, like building features or watching model inputs
Making warehouse data usable by AI tooling; semantic layers, data contracts, or documentation that agents and humans can both read

Responsibilities

Own a scoped domain of dbt models: design, build, test, ship, and monitor them, with a clear point of view on how they should be structured
Turn ambiguous asks from Marketing and Product into scoped work, and talk openly about tradeoffs when the ask and the timeline don't fit together
Write the design doc for the features you own and break the work into pieces teammates can pick up
Add monitoring and alerting to your models so your team catches problems before stakeholders do
Take your turn on our support rotation, debug what breaks, and prevent the repeat
Leave behind runbooks, schema docs, and diagrams that make your work easy for the next person to own
Coach engineers earlier in their careers on the team, in code review and day to day