Founding Engineer (Product / AI)
We're not rebuilding tax software — we're rebuilding the job. AI takes the grunt work so preparers move up to the advice clients actually pay for.
The bet
Tax prep is a huge profession running on decade-old software. Intake, workflow, and delivery have been the industry's most expensive unsolved problem for years: the 40-page PDF organizer that never comes back, hours spent renaming and reordering client documents, a delivery experience that doesn't match the work behind the return.
The technology finally caught up. LLMs can now do the mechanical work that used to need a junior preparer — read the documents, run the client interview, assemble the workpapers — which frees the preparer to do the judgment and advice that's the actual product. Grove is the platform that makes that the default for every firm.
The role
You'd be one of our first engineers, so the surface area is large on purpose. You own features end to end across the Grove app — prior-year, checklist, the client questionnaire, workpapers, delivery — and the AI underneath them: the LLM pipelines for document understanding, the client interview, and workpaper assembly.
This is as much a product role as an engineering one. You'll sit with a tax firm, find the real problem, design the fix, build it, ship it to production, and watch it get used — often in the same week. No handoffs, no specs thrown over a wall, no staging branch nobody sees.
What you'll work on
- Ship full-stack features across the Grove app that real preparers use during real tax seasons.
- Build and harden the LLM pipelines behind document understanding, the adaptive client interview, and workpaper assembly — on messy, real-world tax documents where a single missed form can cost a client thousands.
- Make Grove fast, reliable, and secure enough for tax season, when downtime isn't an option.
- Get on calls with firms, turn what you hear into product, and close the loop yourself.
What we're looking for
Expertise is the floor, not the differentiator. We want engineers who are genuinely excellent at their craft, then pointed at outcomes instead of tickets.
- A strong track record of shipping production software end to end — you've owned things, not just closed them.
- Full-stack range: comfortable across the interface a preparer touches and the backend and data behind it.
- Real fluency with LLMs in production, not demos — you've built things that have to work on inputs you don't control, and you know how to make them reliable.
- You move fast and finish. Speed to you means getting the real version in front of customers sooner, then making it better with what you learn.
- You want to understand the customer's world — a preparer's actual February, the deadlines, the missing documents, the client who never replies. You don't need a tax background; you need to want one.
Nice to have
- You've been early at a startup before — or have always wanted to be.
- Document AI / extraction, RAG, evals, or agentic pipelines.
- Background in tax, accounting, fintech, or another regulated, detail-critical domain.
How we work
- Ship to production. Your work reaches preparers in days, not quarters.
- Talk to customers. Discovery and delivery are part of the job, not another team's.
- Build AI-native. LLMs are why Grove can exist now — in the product and in how we build. We expect you to push that further.
- Built alongside the firms who use it. Every feature ships from a real workflow problem a real preparer is having this week.
Details
Come build with us.
If owning a problem from a preparer's desk to production sounds like your kind of work, we'd like to hear from you.