Analytics Engineer

Want to help everyday Americans invest and build wealth? Financial inequality is increasing, and too many people are getting left behind. At Stash, we’re passionate about democratizing wealth creation through education, advice, and products that help customers achieve greater financial freedom. We also believe in working smarter—leveraging AI and emerging technologies to move faster, operate more efficiently, and focus our time on solving meaningful problems for our customers.

We’re looking for a Senior Analytics Engineer (Technical Level 4) to own and evolve the analytics foundation that powers Stash—our dbt-powered data mart, Looker semantic layer, and the quality systems that keep daily numbers trustworthy for Product, Growth, Finance, and Data Science.

You’ll sit at the intersection of data engineering and data science: production-grade SQL modeling, testing and freshness, clear metric definitions, and close partnership with stakeholders who depend on self-serve data. Company bets (quality growth, Financial Advice, tier packaging, OKR visibility) all run through the mart—if models break, definitions drift, or sources go stale, the business loses trust. Your job is to make that trust durable.

This position operates on a hybrid schedule, requiring you to be on-site in our New York office at least three days per week to work with stakeholders and foster team collaboration.

What you'll do: 

• Own core mart domains end-to-end: Design, build, and maintain production dbt models (bronze → silver → gold patterns in our data mart) for high-priority domains such as subscriptions, promotions/attribution, acquisition, and product usage.
• Raise reliability and quality: Drive down recurring dbt test failures; add meaningful tests; document exceptions; partner on freshness SLAs and alerting so stale or wrong data is caught before Looker, OKRs, or DS models.
• Keep Eng and the mart aligned: Partner with Backend / Product Engineering on instrumentation and schema changes (e.g., service migrations). Reconcile parity, get stakeholder sign-off, and cut over without silent metric breaks.
• Enable Data Science and self-serve: Turn DS modeling requests into governed mart objects (grains, definitions, consumers). Build Looker explores/views and documentation so analysts and PMs can answer questions without waiting on a ticket for every pull.
• Improve ops and performance: Contribute to mart job reliability (retries for transient failures, clear ownership of non-retryable logic failures). Profile and refactor high-cost models when reliability work is on track.