Product Design Engineer

Accelerant is a data-driven risk exchange connecting underwriters of specialty insurance risk with risk capital providers. Accelerant was founded in 2018 by a group of longtime insurance industry executives and technology experts who shared a vision of rebuilding the way risk is exchanged – so that it works better, for everyone. The Accelerant risk exchange does business across more than 20 different countries and 250 specialty products, and we are proud that our insurers have been awarded an AM Best A- (Excellent) rating.

For more information, please visit (Use the "Apply for this Job" box below). .

This role is for someone who helps shape what’s worth building, designs it by building, and owns the result in production. You explore and prototype like a designer, then carry the strong ideas all the way to a shipped front-end, the exploration and the building as one craft rather than two stages with a handoff between them.

That’s possible because of how we work. When the same person designs and builds, there’s no spec to hand off and no rebuild cycle, so the work moves faster and the original intent carries straight through. You work in Product & Technology squads built for the AI era, alongside software engineers who own the systems beneath the experience: they’re accountable for whether it runs, you’re accountable for what it is.

This is not a designer who occasionally prototypes or a front-end engineer who takes tickets.

Most of what gets called “AI in product” today is a button bolted onto an unchanged screen. We’re building the opposite: experiences where the agent does the work and the human reviews, corrects, and authorizes it. Knowing the difference, and being able to design and ship it, is the judgment we’re hiring for.

You’ll join Agentic MGA, the squad that powers risk selection before a policy is bound. The squad gives underwriters and members self serve pre bind analytics, pricing and appetite insight, and intelligence that learns continuously from submissions, bind outcomes, and claims performance.

The surfaces you design carry real underwriting decisions: what to quote, what to decline, what sits inside appetite and what does not. The system learns from what happens after those decisions, so the experience has to make the reasoning legible, not just deliver an answer. Some of this is making dense insurance data clear and usable under time pressure. Some of it is rethinking a workflow entirely now that an agent can carry the load.

You will own these from shaping the problem through to the production front-end, working with software engineers on the systems underneath.

• Explore and prototype to find the right answer. Real design happens here: trying directions, putting things in front of people, following the good ideas and throwing out the rest. Most of what you make is for learning, not shipping, and that’s the point. Production is where the right answer lands, not where it starts
• Carry features from design through to production front-end. The design decision and the code are the same job, in the same hands, so the work keeps its fidelity instead of being lost in a handoff
• Help frame what’s worth building. Talk to underwriters and members, test whether the problem is real, and shape scope alongside your squad rather than waiting for a finished spec
• Build the supervisory surfaces of an agent-native platform: a submission triage queue an underwriter trusts at a glance, a confidence and citation pattern they can act on, an appetite or pricing signal that shows its reasoning
• Use AI tools across the full loop (research, design, prototype, build, test, iterate) as the default way you work, not an occasional accelerator
• Own the invisible craft that…