The AI-native stack for modern underwriting — bespoke to every lending use case. Underwrite borrowers anywhere in the world in minutes. Backed by Y Combinator.
About the role
Credit is the difference between a business that grows and one that doesn't, a family that absorbs a shock and one that doesn't. Kita decides who gets assessed fairly and how fast. That is the scale of what you'd be working on: not a metric on a dashboard, but whether capital reaches people it has never reached before, in five countries and counting.
What they're looking for
- You go deep. You read the papers, you build the thing, and you're usually a step ahead of where the field was last month
- You research and build in the same motion. New capability lands, you've already tested whether it changes what we should ship
- Genuinely curious about the whole company. Our engineers sit in on sales calls, read customer contracts, and work closely with our customers to understand how they make decisions
- You operate clearly in ambiguity. Not tolerate it, operate in it: you can take a vague problem, find the real constraint, and come back with a plan
- Production experience with LLM-based systems and the aftermath, including hallucinations, drift, evals, versioning, and cost
- Strong backend fundamentals. Python and TypeScript are our daily drivers, but how you think matters more than what you've used
More about this role
Credit is the difference between a business that grows and one that doesn't, a family that absorbs a shock and one that doesn't. Kita decides who gets assessed fairly and how fast. That is the scale of what you'd be working on: not a metric on a dashboard, but whether capital reaches people it has never reached before, in five countries and counting.
You'll build the agentic systems that do the underwriting work, harden them until they hold up under enterprise volume and regulatory scrutiny, and work directly with the founders on what gets built next.
- Agentic workflows that run the full assessment: collecting a borrower's file across WhatsApp, Viber, SMS, and email, chasing what's missing, reconciling what doesn't match, running the financial analysis, and producing a decision-ready assessment
- Hardening. Getting these systems from working to reliable is most of the job: evals, failure modes, latency and cost at volume, graceful degradation when a model or an upstream system misbehaves
- Auditability and traceability. Every number we return has to trace back to its source and survive a regulator, an auditor, and a credit committee. This constraint shapes the architecture, it...
Browse similar: AI jobs · AI startup jobs · Startup jobs · Founding team roles · San Francisco Bay Area