Kinro is an autonomous insurance brokerage. Software runs the business: it talks to business owners, prepares quotes, places policies, and services them, with people stepping in only where judgment or licensing requires it. You will build that core. This is a hands-on role for an engineer who has shipped AI products end to end and wants their code to run a real company.
What you'll do
- Build the systems that operate the brokerage: customer conversations, quoting, placement, and policy servicing, designed to run without a human in the loop.
- Own features end to end, from backend services and model integration to the monitoring that proves they can be trusted unattended.
- Push the boundary of what runs autonomously by making our automation faster, safer, and more reliable.
- Work across the stack wherever the leverage is highest, from customer-facing surfaces to core infrastructure.
- Turn every lesson from production into a system that handles the next case without you.
What we're looking for
- You have built AI products or LLM systems end to end in production.
- You think in terms of automation: when a system can act alone, when it should escalate, and how to widen the first category over time.
- You understand the practical tradeoffs of language-model systems: latency, reliability, security, and cost.
- You can reason across the stack: application code, APIs, databases, cloud systems, and operational tooling.
- You have strong product taste and care whether something feels simple, fast, and trustworthy.
Nice to have
- Experience with evaluation systems, model routing, or safety work in production AI.
- Experience in high-ownership environments where engineers talk directly to customers.
- Experience in insurance, fintech, or other regulated industries.