The role
- COMP
- $175K - $350K
- EQUITY
- Highly Competitive Equity
- LOCATION
- New York
- WORKPLACE
- On-site
- EXPERIENCE
- 4 - 12 years
- VISA
- None, Visa transfers
- STACK
- Python, PostgreSQL
- INDUSTRY
- AI, B2B, Finance, Financial Services, Fintech
/roles — ROLE_333
Member of Technical Staff - Applied ML
Fast-growing accounting AI company whose multi-agent systems handle real accounting work in production
The company
Accounting-AI company whose agents do real accounting work end-to-end, used by top firms across client accounting, tax, and audit.
- STAGE
- scale-up
- FUNDING
- $140M+ raised
- TEAM
- 100+ people
- FOUNDED
- 2023
- BACKING
- backed by Khosla Ventures
JD — the work
About the role
In this ML engineering role, you would own projects from scoping to production, taking responsibility for systems that let agents plan, reason, and grade their own output. You set your own plans, define success, run the experiments, and make the call on when a system ships. The role suits people who want to combine research-style experimentation with shipping production systems, instrumenting deeply so that what you build keeps getting smarter. It is on-site in New York.
What you'll do
- Build multi-agent systems that take over genuine accounting workflows end to end
- Set autonomy limits, tool-use rules, and fallback paths that keep agents dependable
- Handle memory and context so agents stay coherent over many steps, and route models to balance latency, cost, and accuracy
- Create online and offline eval pipelines that execute large experiment sweeps without manual work
- Set up golden task sets, labeling approaches, metrics, and regression tracking
- Convert unstructured, messy files into clean data and build retrieval, indexing, and prompt layers around them
- Write clear specs, add validation and guardrails, and communicate what is built, learned, and next
What they're looking for
- Roughly 4 to 12 years of engineering experience building ML-driven systems
- Strong Python and backend skills, with PostgreSQL experience
- Ability to scope projects independently and communicate progress clearly
- A habit of letting experiments and metrics, not intuition, drive decisions