Lead Product Designer - Alt

Job Description

What You'll Own:

• The problem. You write the framing for your surfaces: who's hurting, the evidence, the number that says it's real, the metric that says it worked. AI-assisted, design-reviewed, never skipped. If nobody can state the problem in a sentence, it isn't ready to be solved.

• The craft. How completely the solution solves the problem, all the way to delight. It runs the full arc: reading the problem so precisely that the solution feels obvious in hindsight, shaping the flow so the user never has to think about the tool. It takes a user from "this worked" to "I want to show someone." That's still made by a person with taste and reps. We can generate a hundred options in an hour; you're here because you can make one of them solve the problem so well it delights.

• The solution space. Multiple real directions, converged against your surface's written bar. You'll pressure-test, extend, and validate with users — not pick the first thing that renders.

• The experience decisions. When a thing appears, what it feels like in the user's actual state, what tone it takes, what gets removed to make room, and what it does to neighboring screens. On a platform where people wire five figures based on what a screen makes them feel, these decisions are the product.

• The words. Copy is design here. The sentence that explains why an offer is what it is does more work than the layout around it. You write it.

• Trust, specifically. How much confidence we assert, when we show uncertainty, what the platform states versus suggests. As more of what we show a user is model-generated, this becomes the highest-stakes surface in the company, and Design has sign-off on it.

• The crafted base. The quality foundation your team iterates on, and its survival through implementation. You pair with engineers at ship time; built-versus-designed fidelity is yours.

• Your metrics. Instrumented before launch. You run your own readouts and kill your own misses.

• The research. Customer conversations every week, CX tickets and session replays mined as a habit, lightweight tests run by you, at the pace of the work. Your frames cite evidence because you went and got it.

• The prototype. You build. Working prototypes in code, not click-throughs, are how ideas get evaluated here, because a thing you can touch settles an argument in an afternoon that a doc would run for two weeks.

How You'll Use AI Here:

We don't want to hear that you use Claude every day. Everyone does. We want to know what you've built with it. Concretely, in this role:

• Get to a clickable prototype of a new flow in a day — real card data, simulated model output, testable interactions — so the team argues about a real thing instead of a doc

• Write a Claude skill that holds your surface's written bar, so every frame gets checked against it before it reaches review, and the bar stops depending on who's in the room

• Mine CX tickets and session replays at volume to find the pattern you'd never catch by hand — the phrasing that precedes an abandoned submission, the screen people revisit three times before they bail

• Turn a customer call into a synthesis the team can act on the same afternoon, not a deck three weeks later

• Generate and kill twenty directions before you commit to one, and be able to say what each one taught you

• Put a prototype in front of actual high-value sellers and come to roadmap planning with behavioral data — "version B moved stated acceptance likelihood, here's the prototype, here's the number"

And have a view on the other side of it: where AI output is actively wrong for this work, and what you refuse to hand to a model. We ask.