Design Engineer

Socket.dev; San Francisco (CA); On-site; USD 120,000 - 160,000; Full time

Job summary

Gooseworks is building AI-driven creative for consumer brands. You will own design end-to-end, shipping features from product decision through frontend live release. You’ll craft a design language that agents can follow and build a lightweight design system for agent-powered workflows.

You’ll ship your own frontend using React, Next.js and Tailwind, with a strong emphasis on real, working code and practical UX judgment.

Qualifications

• Real design work, shipped and live — not a case-study deck.
• You ship your own frontend with React, Next.js, Tailwind.
• You vibe code effectively, leveraging AI agents in design.
• CS background or technical design background is a plus.

Responsibilities

• Design end-to-end user experiences and raise the design bar for our product and brand.
• Develop the design language for AI agents to follow.
• Build features end-to-end from product decision to frontend shipped.
• Create mini-apps and lead magnets, 1-2 day builds, end-to-end.

Skills

Product design

UX judgment

Design system

AI agent design

Design language

Education

CS background

Tools

React

Next.js

Tailwind

Job description

Growth is the hardest problem every company has. Not building the product — getting it in front of people, over and over, in a way that works. Today that means an army of humans making creative, testing it, reading the numbers, and doing it again next week. It doesn't scale, and it's the single biggest tax on every consumer brand in the world.

We think that changes. The billion-dollar companies of the next decade won't scale linearly with headcount — they'll be run by fleets of AI agents, and the org chart will look more like an architecture diagram than a hierarchy.

We're starting with the ad account. Gooseworks builds AI agents that make and run creative for consumer brands — ads, UGC, organic — end to end. Very early, but the pull is real:

• 200+ paying users
• 30,000+ ads generated by our agents in the last 2 months

We're a YC company (W23), based in San Francisco, and we run the company on agent harnesses.

About the role

You'd own design at Gooseworks. All of it.

You design it and you ship it — small and medium features end to end, from the product decision through the frontend that goes live.

Not figma mockups. Actual, functional PRs, working and tested.

There's a second half to this job that's less usual, and it's the interesting one. Most of our product gets built by agents. That means design consistency isn't enforced by a person reviewing pull requests — it's enforced by whatever we've written down for the agents to follow. So part of your job is building a design language and a component set that _an agent can actually consume and get right_.

What you'll work on

• Product design. You'll design end-to-end user experiences, and raise the design bar for our product and brand.
• The design language, for agents. Teaching AI agents how to design really well.
• Building features, end to end. Product decision, design call, frontend, shipped.
• Mini-apps and lead magnets. Small, sharp, fast things we put in front of people. 1-2 day builds, end-to-end.

We're looking for

• You're a designer first. Real design work, shipped and live — not a case-study deck. We'll open the thing.
• You ship your own frontend. React, Next.js, Tailwind. You don't need to be a systems engineer, but you do need to be able to get your work in front of users yourself.
• You vibe code effectively. Your output should be visibly agent-leveraged — you drive Claude Code or Codex, you build skills and harnesses to make them better, and you push them well past autocomplete. It's the highest-signal thing about a candidate for us, because it's how the whole company operates.
• Product and UX judgment. You can argue about what to build and why it should feel a particular way.
• A technical foundation. Studied CS, worked as an engineer, or otherwise came at design from the technical side. Not required, but a strong plus point.
• Nice to have: founding-designer experience at an early startup, motion or video ability, prior work in an AI-native product.

If you design things that look considered, ship them yourself, and have opinions about how agents should be taught to build — send us the link.