(Hands-on role, Stage: Pre-Seed)

Location: NYC or San Francisco, USA

Context

At resfi.ai we represent responsible finance.

We are building the first truly ‘agentic neobank’ that will be a true customer advocate not only giving but also implementing advice automatically for consumers on their day-to-day financial needs by leveraging AI and behavioral finance.

We are building an AI-native banking, lending, and affiliated services from the ground up.

We believe the next generation of financial products will be autonomous, not rules-based. An everyday consumer will not have to worry about or actively manage their finances. We are building an autonomous personal CFO that will optimize every element of your personal finances - across idle cash savings, debt, budgeting, expenses, investments, and taxation.

Founder, Asheet Mehta, is driven by a commitment to helping consumers achieve stronger financial outcomes. Before ResFi, he was a Senior Partner at McKinsey & Company and served as Co-Head of its Global Financial Services Practice.

This role exists to begin by building the intelligence layer of the company — systems that reason, decide, and improve over time, while remaining safe and explainable.

There is no existing stack. You will build it from zero.

You will be one of the first technical leaders evaluating tech vendors, writing real code, making architecture decisions, and shaping the culture.

What You’ll Do

• Influence all technical decision-making from technical vendor selection or architectural decisions
• Design and build early AI and data architecture that:
  • is robust end-to-end for autonomous decisioning to automatic actions (think a team of financial AI agents for personal finance managed by an autonomous AI CFO)
  • Build agentic and decisioning workflows where appropriate (from customer-facing to internal ops - from ledger automation and compliance documentation to voice AI for customer servicing and personalized messaging)
  • Remain explainable, auditable, and regulatory compliant
• Evaluate trade-offs and decide when to use:
  • LLMs
  • Traditional ML
  • Deterministic logic
• Build evaluation and monitoring from day one
• Rapidly prototype, ship, discard, and rebuild
  • Bias for action and aggressive experimentation is expected
• Define clean interfaces between AI systems and financial core systems