VP, Product

Insurance underwriting runs on human judgment, manual workflows, and knowledge locked in someone's head. It breaks when that person leaves and doesn't scale when the book doubles. Sixfold is building AI agents that understand risk the way a 20-year veteran underwriter does. Not summaries or flags, but systems that make real decisions on billions in premium for some of the world's largest carriers. AI is the product, the workflow, and the moat. If you think in systems, move fluidly from research to pixels, and are energized by designing AI-native experiences where the best practices haven't been written yet — this role was made for you.

Who We Are

🏔 Product: Sixfold is the AI brain for underwriting. Our platform handles submissions at scale, assessing risk against appetite, surfacing cited insights, and deploying AI agents to handle research, referrals, and triage so underwriters can focus on building profitable portfolios.

💪 Team: A focused team of insurance operators, AI researchers, and engineers who love hard problems. We've pushed the boundaries of what AI can do with unstructured data — from teaching AI to understand that a bakery in Florida faces different risks than one in Montana, to knowing when a manufacturing company's pivot from toys to medical devices fundamentally changes their risk profile. We ship fast and stay close to customers.

📈 Traction: Processing billions in premium for some of the world's largest carriers. We know AI. We know underwriting. We're building what comes next.

About the Role

You'll inherit a strong product and a capable team, but that's the starting point, not the job. The job is figuring out what underwriting becomes once AI stops assisting and starts deciding, and getting there faster than the market expects. You prototype to find out, not to perform speed. You're drawn to this problem because the underwriter of the future is one of the more interesting design problems in software right now.

In this role, you will:

Set the direction for what AI-native underwriting looks like at Sixfold, and hold a 12-month theme even as the tactics under it change frequently
Build the first version of new ideas yourself: a prototype, a working demo, a lighthouse customer test, before asking the team to build the tenth
Stay current enough on AI’s leading edge to have a real opinion on where things are going, not just what’s already shipped
Represent Sixfold's view of where underwriting is headed with customers, partners, and the industry, so people outside the company measure their own thinking against yours
Take ownership of the existing product and team and raise the bar on how they operate, without mistaking "it's not broken" for "it's finished"
Partner closely with design and engineering so prototyping speed doesn't turn into technical debt. Speed and guardrails are both your job, not competing ones

What Success Looks Like

Within three months, we expect you to have a visceral understanding of our risk-reasoning AI. If you're still relying on our current product docs instead of having your own opinion on where the edges are broken, we’ve missed the mark. By the end of the first quarter, our engineering team should be moving faster, not because you managed them, but because you provided more clarity to them

About You

You think about the underwriter of the future the way some people think about a chess position. You can't stop turning it over, and you have hot takes about what's underrated
You said you built something yourself, you mean you wrote the prompt chains, ran the eval, sat with the customer, not that you wrote the spec and someone else built it
You took from zero to production, and you know the ARR, retention, or adoption number it moved
Facing a new problem, your instinct is to build something small and get it in front of a real customer rather than write a spec. You find the lighthouse customer, get honest feedback, and iterate from there
You can hold a clear 12-month theme for the product. The destination doesn't move even if the path there does. You just don't confuse that with a fixed roadmap of mechanisms, since you expect to rebuild those every few months as reality changes