About the role
We ship UI fast: a team of under ten plus AI coding agents that write production pull requests. Speed is not our problem. Coherence is, and this role owns it: the design tokens, the component library, the frontend and conversational interface patterns that make everything we ship, human-written or agent-written, feel like one product.
We’re looking for a Design Engineer to raise the bar on interaction quality, motion, and polish across our product. This role sits at the intersection of design and engineering: you'll work directly with the executive team, pricing architects and engineers to ship new features that feel crafted, consistent, and human.
What you'll do
- Own the design system: design tokens, the shadcn/ui-based component library, and Tailwind conventions across our product suite.
- Shape the brand identity of our conversational agents and bots within our customer’s communication platforms like Slack and Teams
- Define the React Query, React Hook Form, and Zod patterns that engineers and AI coding agents both follow, so agent-written pull requests land consistent by default. The design system serves as the guardrail.
- Enforce accessibility. WCAG contrast rules are written into our brand guidelines, and you make the product honor them.
- Own frontend tooling and quality: Biome configuration, Playwright E2E coverage, and frontend performance.
- Partner directly with the founders on how the brand shows up in product.
- Keep the frontend what our architecture says it is: a thin, replaceable layer over the versioned API. Your job is to make thin also mean polished.
Your first 90 days
First 30 Days: Audit & Workflow Calibration
- Complete a comprehensive audit of existing tokens, Tailwind conventions, and components across the product to map out top UI fragmentations.
- Pick the highest-friction UI inconsistencies and ship your first token and component cleanup across our product suite modules.
- Onboard into our AI-native SDLC using Claude Code and Cursor to analyze how AI agents currently write frontend code and where they drift from design intent.
First 60 Days: Codifying Guardrails & Enforcing Patterns
- Formally establish the design tokens and shadcn/ui component library, systematically eliminating one-off components from core product flows.
- Define and enforce React Query, React Hook Form, Zod, and Tailwind patterns in Biome lint rules and reference implementations so agent-written pull requests land consistent by default.
- Codify WCAG contrast requirements and brand guidelines directly into code rules, executing a formal accessibility audit across core flows.
First 90 Days: System Ownership & Cross-Platform Reach
- Reach a state where agent-written pull requests pass review without pattern or styling corrections, with the design system serving as the automated guardrail.
- Establish Playwright E2E regression coverage for key user flows through proposition management, personas, and surveys.
- Define and own component patterns and visual guidelines for our conversational agents and bots within platforms like Slack and Teams.
What we're looking for
- 5+ years building product interfaces, with a portfolio of shipped features you can point to that rivals Linear/Vercel-like polish
- You have owned a design system or component library in production, as its primary owner. Public repo, portfolio, or a detailed walkthrough of proprietary work all count.
- You care deeply about typography, spatial rhythm, motion, and tactile micro-interactions. You instantly spot when a screen is 2px off or a transition feels sluggish ßand you refine it directly in code
- Deep expertise in React, strict TypeScript, Tailwind CSS, and component primitives (shadcn/ui, Radix UI). You know how to keep a typed codebase honest, flexible, and scalable.
- Taste plus systems thinking: You don't just tweak a screen once; you translate design intent into design tokens, component abstractions, and lint rules. You turn your taste into automated guardrails so that code written by humans or AI agents stays coherent by default.
- Comfortable being the sole bar-raiser for user experience, interface craft, and frontend standards across the entire company.
- You can show how AI coding tools fit into your work today. We weigh that over where you studied or your previous role.
- Work authorization: You must be authorized to work in the US. We're unable to sponsor visas at this time.
Nice to have
- Experience building non-standard AI interfaces (e.g., streaming responses, generative canvas views, human-in-the-loop workflows) using the Vercel AI SDK or custom state machines.
- A deep toolkit for animation and UI state transitions (e.g., Framer Motion, Radix transitions) that make interfaces feel fast, tactile, and alive.
- Ability to refine visual details, layout, and hierarchy directly in code without waiting for high-fidelity Figma handoffs.
- Experience working alongside AI coding agents such as Claude Code or Cursor specifically writing custom rules (eg .cursorrules, claude.md, custom Biome/ESLint rules) to guide component generation and styling.
- Experience designing and shipping conversational interfaces or chatbots that delight B2B end-users.
- Familiarity with Slack Block Kit, Microsoft Teams Adaptive Cards, or CLI design patterns, where visual real estate is constrained and typography/structure carry the brand identity.
Our stack
Client: Typescript, React, Vercel Chat SDK
Server: Python, FastAPI
Data: Mongo, Atlas
Infra: GCP, Pulumi, Cloudflare Pages
AI: FastMCP, Claude Code, Cursor, Vercel Eve
Security: SOC2 Type 1/Type 2, GDPR compliant, EU and US data residency
How we work
We are under ten people, and everyone ships and talks to customers. Engineers are product engineers: you own outcomes, scope your own work, and demo every week.
We work in-person 5 days a week at our downtown San Francisco office. We believe the best products get built when people are in the same room. Fast feedback loops, real collaboration, and a team that genuinely enjoys working together. If that's how you do your best work, you'll fit right in.
There is no separate PM layer. Our SDLC is AI-native as we move to a robust software factory. Claude Code and Cursor are standard kit, and LLM agents write and review pull requests behind automated review gates.
We keep pull requests small, around 500 lines, because the research on small batches holds up and we act on it.
The architecture rules are short and enforced: the API is the product so it evolves additively, business logic stays server-side, and writes are audited by default for compliance. Here, the design system is how agent-written UI stays coherent by default.
What we offer
- Compensation: $180,000 - $220,000
Plus meaningful equity through our employee stock option plan (ESOP). We aim to be highly competitive on cash and generous on equity. Joining at this stage means a real stake in what we build together.
This compensation and benefits information is based on our good faith estimate for this position as of the date of publication and may be modified in the future. Employees based outside of the US will receive a different benefits package. The level of pay within the range will depend on a variety of job-related factors, including where you place on our internal performance ladders, which is based on factors including past work experience, relevant education, and performance on our interviews or in a work trial. - Benefits: A market-conform package, including health coverage, pension/retirement provision, generous paid time off, paid parental leave.
- Direct customer contact from early on: you hear how your recommendations land.
- Real influence on the architecture while the big decisions are still open. The AI layer is young, and you set its shape.
- A team where your impact is visible.
- An opportunity to be part of a fast-growing company.
- Daily work with, and real influence over, a cutting-edge AI pricing engine.
- Continuous exposure to the newest AI tools and techniques.
How to apply
Send a link or a short written walkthrough of a design system you built (proprietary is fine), plus your take on how to maintain brand fidelity across different interfaces (Web UI to conversational interfaces to terminal interfaces). If you have public work, link it.
Don’t check every single box? Apply anyway.
We set a high bar, but we also know that exceptional design engineers rarely fit neatly into a job spec. They often come from diverse and non-traditional backgrounds. If you don’t meet 100% of the qualifications above, but you have high visual taste, deep code craft, and are energized by setting the UI standard for an AI-native product, we want to hear from you. We value trajectory, high agency, and shipped work over perfect checklist compliance.