Software Engineer (Senior)

Maximor AI•New York City, NY

9h•Onsite

About The Position

Maximor is an AI agent platform for finance. Our agents connect to ERPs, CRMs, billing, payroll, and banks and automate the manual accounting work that consumes finance teams every month: reconciliation, journal entries, contract parsing, revenue recognition, close checklists, and audit-ready reporting. Finance teams review only the exceptions — every output is traceable and audit-ready. We've raised $9M led by Foundation Capital , with BoldCap, Gaia Ventures, Aravind Srinivas (CEO, Perplexity), and finance leaders from Zuora, Zoom, Ramp, Gusto, MongoDB, and the Big 4 — one of the largest seed rounds in the accounting category. The Day-to-Day Senior engineers own an accounting module — close, revenue, cash, reporting, accruals, or fixed assets. You join a pod of 2–3 engineers and ship the whole thing: backend services, agent workflows, and the surfaces accountants use. No platform team builds for you. No PM hands you specs. Problems Worth Your Brain What's the right shape of a sub-ledger primitive that every agent in the company writes through flexible enough to model any accounting workflow, rigid enough to keep an audit trail? How do you build an agent platform where pods ship reliable Audit-Ready Agents in days, not months — without each pod re-deriving the same eval, guardrail, and observability work? How do you encode auditability as a property of the system itself, not as something each pod has to remember to add? How do you build a close orchestration layer that's flexible enough for 1,000 different close calendars but rigid enough that an auditor can trust it? What does "engineering excellence" mean for non-deterministic systems in regulated environments — and how do you make it teachable to the next engineer who joins? How do you make sure the architectural decisions made in 2026 don't trap the company in 2029? We don't have these figured out. That's most of the appeal. A Few Strong Opinions The engineer who can't operate AI agents fluently is becoming obsolete, fast. Fluency with Claude, Cursor, and internal agent tooling is a second cortex. "Backend vs. frontend" is dissolving. The agents do the typing. The constraint is product judgment, system design, and the ability to close the loop from problem to shipped feature. Our engineers ship full-stack when the work calls for it. The pod is the unit of leverage. Two engineers who can hold an entire module in their heads ship more than ten engineers who each own a slice. Specialization across pods, generalist within them. Our engineers raise the ceiling for every pod. The subject matter expert + engineer loop is the moat. Engineers who sit with controllers and build from what they see compound faster than the ones who don't. What Great Looks Like Here You can learn an accounting workflow in an hour and have it in code by the end of the day. A controller walks you through a prepaid amortization across 80 entities; you spot the edge cases she didn't think to mention. The single most important skill on the team. You're a current or future founder. You scope your own work, think about the customer, own your decisions. You solve problems end to end. The team is split vertically, so every engineer owns a part of the product and makes decisions across the LLM pipeline, infrastructure, backend, and UX. You care about getting it right. A 100% solution beats an 80% one. When something breaks, you go to root cause. You operate AI coding agents at a high level. You have opinions about which agent to use for which task. You've shipped real production code through them.

Requirements

5+ years of software engineering, mostly backend or infrastructure.
Strong backend chops in a modern language (Python, Go, Java, Rust, or similar).
Built hard backend systems (distributed pipelines, transactional or ledger systems, integration platforms, or workflow engines).
Worked at an early-stage startup (pre-seed through Series B).

Nice To Haves

Worked at an accounting, fintech, or ERP-adjacent SaaS startup.
Built and deployed AI agents in production.
Worked with the modern data stack (dbt, Snowflake, Fivetran) or with sub-ledger / general ledger systems.
Built eval frameworks or LLM observability tooling.

Responsibilities

Own an accounting module (close, revenue, cash, reporting, accruals, or fixed assets).
Join a pod of 2-3 engineers and ship backend services, agent workflows, and accountant-facing surfaces.
Operate AI coding agents at a high level and ship production code through them.
Scope own work, think about the customer, and own decisions.
Solve problems end-to-end, making decisions across the LLM pipeline, infrastructure, backend, and UX.
Go to root cause when something breaks.