Staff Product Designer, Guest & Discovery

Company

Collaborative Boating operates Boatsetter and GetMyBoat, the two largest marketplaces for getting on the water, now converging into one platform. Guests arrive wanting something specific — a fishing trip, a sunset cruise, a birthday on a pontoon — and we connect them to the captains, operators, and owners who can make it happen. Finding and booking a day outside should be as easy and intuitive as booking a flight.

The problem

Search is changing under us. Guests arrive with intent expressed in natural language, increasingly through an AI assistant instead of a search box, and increasingly on a surface we don't own. Answering that with a chatbot is the obvious move and the wrong one: choosing between boats is a parallel, visual, comparison task, and a conversation is sequential. The work is holding both — a visual discovery experience built for comparison, and a compact, structured expression of a listing that holds up when a third-party assistant renders it in a layout we didn't design.

The center of gravity is the search results card. It's the smallest unit of persuasion we have: a handful of facts, ordered, that a guest uses to decide whether to look closer. Which facts earn a place, in what order, and how that adapts from "boat rental Miami" to "something fun for eight people Saturday" — that's the problem. It's an information design problem before it's a visual one.

What you'll own

• The end-to-end guest discovery experience — search, filtering and refinement, results, map and list, the listing page, and the paths in from organic search and AI assistants. Coherence across both brands and web plus mobile is yours.
• The information hierarchy of the results card, defended with evidence, across guest intents and every state the data actually produces: missing photos, no reviews yet, a price range instead of a price, partial availability.
• Interfaces over a system that infers. Make what it inferred legible, make correcting it easy, and reduce verification labor rather than creating it — a guest who has to double-check our answer has been handed our work.
• Surfaces we don't render. With engineering, shape the structured expression of an experience, an offer, and an operator so it survives being summarized, quoted, or laid out by someone else's assistant.
• Continuous discovery with your pod — a Product Manager and an engineering tech lead. Choose the cheapest artifact that answers the open question; spend fidelity when a decision is actually being committed.

How we work

Five things worth knowing before you apply. They aren't universal, and here they aren't negotiable.

Problem, then system, then experience

We frame the problem, then settle the shape of the system — entities, states, edge cases — and only then design the experience within it. Your rough exploration runs alongside the system design and improves it; committed fidelity comes later. In practice that means days of useful work before there's a screen worth showing, and arguing about data models with your tech lead while the design is still loose. If you need a settled brief before you can start, this will be frustrating.

Discovery is ongoing, not a handed off spec

You talk to guests and operators directly. You bring opportunities forward rather than receiving them. There's no research team handing you findings and no Product Manager writing you requirements — the pod owns the outcome and you're one third of it. We believe the commercial results follow from understanding these customers better than anyone else does, which is why the understanding isn't delegated to a function.

We ask what moved, not what shipped

You define the signal before release and watch it after. Hitting the number you set is the floor, not the finish — the work worth remembering beat it, and we'd rather you aim at that and miss than aim at the floor and land. Sometimes the answer is that your design didn't work, you say so plainly, and we kill it. If measurement feels like a threat to the craft rather than part of it, this will be a hard place to work.

We ship in days, not quarters

Work goes live the day it's ready — no launch calendar, no approval chain. The pod decides and it goes out. Fast here means small and finished, not big and rough: a narrow slice with its states designed and its accessibility handled, in front of real guests this week, then the next slice. The quality bar doesn't move; the amount of surface you hold before releasing does. What you give up is the reveal — your work is live while it's still incomplete in aggregate, and you correct in the open rather than perfecting behind a curtain. Designers who need to control the whole surface before anyone sees it will find this hard.

We cross lines, and we bet on each other

Titles and team boundaries carry little weight here. You'll be inside the domain model with an engineer, in a support queue reading what operators actually complain about, in a pricing conversation that isn't nominally yours. Ask for help early and give it generously. The standing assumption is that the person next to you is good and worth backing before they've proven it to you — and we expect the same read on you.

We work inside the design system by default; a new pattern has to earn itself against what already exists. WCAG 2.1 AA is an accepted platform standard, not a per-feature negotiation.

What you need to have done

• Held substantial ownership of search, browse, or discovery in a two-sided marketplace — relevance surfaces, faceted filtering, availability and pricing display, map plus list, comparison.
• Designed information you can defend. Walk us through a dense comparison surface you built: why each element is there, why it's ordered that way, what you removed, and what removing it cost.
• Designed over a system that infers, ranks, or predicts — communicating state and confidence in real time, and designing the correction path for when the system is wrong.
• Formed a point of view on search in an AI-driven world, including where conversation doesn't belong. We'd rather you disagree with us than agree quickly.

What also helps

Fluency with entities, states, and contracts alongside your tech lead. Restraint and typographic craft in high-information surfaces. AI tools in your own practice. Search relevance literacy, structured data and schema work, third-party assistant surfaces, or travel, hospitality, and local-services search at scale are each a plus rather than a requirement.

Your portfolio

Send work, not a slide deck about process. Two or three cases at depth. For at least one: the problem and how you knew it was real, what you tried and abandoned, the hierarchy decisions and their rationale, the states you designed for, what shipped, and what it changed. Include something that didn't work. Roughness is fine — we care about the reasoning.

Practicals

Remote, with high overlap in United States Eastern hours. You'll join a five-person design team, report to our Director of Design, and be the dedicated designer in this pod. Relevant degrees in design, HCI, or information architecture are welcome and not required — portfolio and shipped, measured work outweigh credentials.