Founding Applied AI Engineer

About the company

Our client is a seed-stage, venture-backed AI startup in New York building agent software that automates accounting work — month-end close, bookkeeping and tax — for accounting firms. The pitch its founders use is an editor for accountants: a vertical agent orchestration platform rather than another dashboard. Founded in 2025, around ten people, $3.5M raised. The founding team pairs a decade inside a Big Four firm with a CTO who has built accounting products at two previous companies, so the domain is understood from the inside rather than guessed at.

The role

You would be a founding engineer, reporting to the CTO, architecting and building the production AI agent systems that automate real accounting workflows: month-end close, bookkeeping, tax and audit. You own systems end to end — prototypes, evaluations, deployment, observability, and the iteration loop with actual accountants using the thing. A large part of the judgement is architectural: deciding where deterministic code is correct, where an AI-assisted workflow fits, and where an autonomous agent is genuinely the right answer, in a domain where a wrong number is not an acceptable outcome. The work is roughly half defined product features and half open-ended AI work, and 80 to 100% of it is greenfield. You would work directly with the founders and with the accounting-firm operators who use the product.

What you'll do

• Architect and build production AI agent systems that automate accounting workflows — month-end close, bookkeeping, tax and audit
• Own systems end to end: prototypes, evaluations, deployment, observability and iteration with the accountants using them
• Decide where the boundary sits between deterministic code, AI-assisted workflows and autonomous agents
• Work directly with the founders and with accounting-firm operators
• Integrate AI into the product experience rather than bolting it on at the edges
• Build mostly greenfield — 80 to 100% of this work does not exist yet

Nice to have

• Accounting, finance, or another regulated or document-heavy domain
• For candidates at the junior end of the range, a degree from a top-tier university
• Experience building evaluation harnesses for AI systems that have to be right rather than plausible

Tech stack

LLM agents, retrieval and document processing at the centre of the product, with the orchestration layer that holds them together, plus the evaluation and observability tooling needed to trust an agent with someone's books. The role is applied AI engineering rather than research: production systems, real users, and the deterministic-versus-agentic boundary decided case by case.

Compensation & logistics

• $200k–$300k base plus meaningful seed equity, with medical insurance
• On-site in New York, with relocation support flexible for the right candidate
• The client can support a visa transfer, such as OPT or an H-1B transfer, but will not file a new, first-time visa sponsorship
• Two hires on this role, both as founding engineers
• Interview process: a 30-minute intro call with the CEO on your applied AI background and production ownership, a 30-minute culture-fit conversation, then a one-hour technical session with the CTO and founding team on agent architecture, orchestration, retrieval and production AI, which sets up a work trial
• Full-time