Hyphenate
Software Engineer (New Grad)
New York City · On-site · Entry level · Full-time · Posted Aug 25, 2026
- Experience
- At least one substantial software engineering or AI engineering internship; finishing a bachelor's or master's degree, or graduated within the past year.
- Education
- Bachelor's or master's degree; computer science, machine learning, or AI preferred, with adjacent quantitative degrees welcome if supported by relevant work.
- Visa sponsorship
- Visa status is not a filter; the company supports OPT and longer-term status.
About the job
Builds audit-ready AI agents and finance software that automate workflows across the CFO office, from data integration and reasoning to product features and updates to financial systems.
What you'll do
- Own end-to-end engineering work within a finance-domain pod, including agent context, prompts, tools, evaluations, guardrails, and product features.
- Build and improve AI agent systems, including verification, observability, durable workflows, and safe, traceable updates to systems of record.
- Work directly with controllers, accountants, and CFOs to understand finance workflows and turn them into software.
What they ask for
- At least one substantial software engineering or AI engineering internship.
- Hands-on experience building AI agents through an internship, research, coursework, or a serious side project.
- Fluency with coding agents such as Claude Code, Cursor, or equivalent, with production code shipped through them.
- Strong fundamentals in a modern programming language; Python is the company's stack.
Nice to have
- Research in AI, ML, or systems; teaching a serious systems, ML, or compilers course; or relevant open-source contributions.
- Experience building evals, verification, or observability for non-deterministic systems; startup experience is also noted as a differentiator.
Tools
Python, C++, Java, Go, Rust, Claude Code, Cursor
AI focus
Build audit-ready AI agents for finance workflows, including context and prompt design, tool use, evaluations, guardrails, verification, observability, and systems that explain decisions and escalate uncertainty.
Model-written summary · the original posting is authoritative