Software Engineer (New Grad)

Maximor AI•New York, NY

21h•Onsite

About The Position

Maximor is building the AI operating system for the CFO office. Our Audit-Ready AI Agents connect to a company's existing finance stack—ERPs, banks, billing, payroll, CRM, contracts, spreadsheets, email, and Slack—and automate the work behind the entire order-to-cash process, record-to-report process, treasury management, financial reporting, and audit readiness. The goal is for finance teams to review exceptions while AI does the rest. Maximor's differentiator is its Unified Finance Context, a financial understanding layer that captures transactions, policies, contracts, historical decisions, and accounting judgment. On top of this layer sit Audit-Ready AI Agents capable of reasoning, explaining decisions, escalating uncertainty, and continuous improvement. The company has raised $9M from investors including Foundation Capital, BoldCap, Gaia Ventures, and prominent finance leaders.

Requirements

Finishing a bachelor's or master's degree, or graduated within the past year.
Computer science, machine learning, or AI strongly preferred; adjacent quantitative degrees welcome if backed by work.
At least one substantial software engineering or AI engineering internship.
Early-stage and AI-native startups count for more.
Hands-on experience building with AI agents (internship, research, coursework, or side project).
Real fluency with coding agents (Claude Code, Cursor, or equivalent).
Strong fundamentals in Python; if depth is in C++, Java, Go, or Rust, be ready to ramp fast.
Something substantial to walk through in depth (side project, research prototype, open-source contribution, hackathon build).

Nice To Haves

Research in AI, ML, or systems (published or in progress at NeurIPS, ICML, ICLR, ACL, or similar).
TA'd or course-assisted a serious systems, ML, or compilers course.
Shipped a side project that real people actually used.
Contributed to open source that other people depend on.
Built evals, verification, or observability for something non-deterministic and caught a bug before a human did.
Interned at a pre-seed through Series B startup and thrived in ambiguity.
Done well in competitive programming or olympiads (ICPC, IMO, IPhO, Putnam, or similar).
Curiosity about fintech, accounting, or how money actually moves.

Responsibilities

Own a finance domain end-to-end within a pod of 2-3 engineers, including context, prompts, tools, evals, guardrails, and surrounding product features.
Scope and ship your own work by the end of the first quarter.
Collaborate directly with controllers, accountants, and CFOs to understand their workflows and build agent systems to replace them.
Build AI agents that finance teams and auditors can trust, focusing on verification, guardrails, observability, and evaluation systems for non-deterministic AI.
Engineer financial context over transactions, ledgers, contracts, policies, and historical decisions.
Ingest, normalize, and reconcile data from various enterprise systems (ERPs, banks, payroll, billing platforms, CRMs, email).
Build agents that improve over time by explaining reasoning, escalating uncertainty, and learning from human corrections.
Orchestrate durable AI workflows across stateful enterprise systems.
Safely write back to systems of record with idempotent, audit-ready updates and full traceability.
Operate AI agents fluently using tools like Claude Code and Cursor.
Ship full-stack when the work requires it, as the distinction between backend and frontend is dissolving.
Collaborate within a pod as the primary unit of leverage, holding an entire module in your head.
Engage in the accountant + engineer loop, building directly from customer interactions.
Translate finance workflows into software.
Think like an owner, scoping your own work, considering the customer's perspective, owning decisions, and driving outcomes.
Solve problems end-to-end, making decisions across the LLM pipeline, infrastructure, backend, and UX.
Dig deep to understand the root cause of issues when something breaks.
Use coding agents fluently and build product-level agents (prompt and context design, tool use, evals).
Communicate clearly, ask good questions, share progress proactively, and raise issues early.