Generalist data engineer to own ingestion and analytics — independently, AI-first, in a fast-moving startup.
We're a small data function inside a tech company at a real inflection point. This role takes ownership of two foundational projects: robust client-data ingestion, and giving the business a live view of how the Forest platform is performing. You'll work independently across teams, with AI as a core part of how you build.
See the roleWe're hiring for mindset as much as for skills. The right person can ramp fast in a small team and own outcomes without needing the work pre-scoped for them.
You can take a vague outcome and turn it into shipped work — scoping, deciding architecture, and building without waiting for direction. Most of what you'll work on starts as a problem, not a spec.
Strong Python and SQL, comfortable with Databricks and Postgres. You've built ingestion pipelines end-to-end and have opinions on how to do it well — deduplication, multi-source, time-to-onboard.
You build with AI tools as part of your default workflow — most coding done through prompt-driven flows like Claude Code rather than direct edits. You know what these tools are good for and where they're not.
A small data and engineering crew. You'd embed in cross-functional teams alongside data and software engineers — supported by, but not dependent on, the people below.
In your first six months you'll work across two priority projects: client data ingestion, and building out Forest analytics for the business. You'll be the dedicated data engineer for these — making your own architecture calls and shipping with AI tooling as a default.
Most days you're working independently on whichever of the priority projects is most urgent. You'll be embedded in a small cross-functional team alongside other data engineers (Jerry, Lily) and software engineers from the Forest and Engage sides — you can ask anyone, but no one is going to spoon-feed you the work. You'll spend most of your coding time in Claude Code, prompt-driven rather than typing edits, and regularly be making your own calls on architecture and scope. Quick syncs with Matt or Jacob keep things moving; the rest of the day is you, the problem, and the build.
Designed to be honest about how you'll actually work — not a series of hoops. Four stages, no take-homes.