2022 — 2026 · Design Lead / UX Engineer
Wolf
The JobSeeker app that accelerated a B2B staffing marketplace
Four years as Design Lead on a B2B staffing marketplace. I designed the JobSeeker app from scratch — the one that unblocked product growth — defined the component system, and shipped the AI-assisted job-request builder.
- Company
- Wolf Inc.
- Team
- Sole Product Designer
- Platform
- Web · iOS · Android
- Stack
- Figma · React · JavaScript · Design Systems · AI tooling
The product
Three surfaces, one ecosystem
Wolf is a B2B SaaS that builds custom marketplace-style platforms for staffing companies. The ecosystem splits into three connected surfaces: a JobSeeker app — the workers who search and apply for jobs, filtering by location and job type — a Client platform — restaurants, clinics, hospitals, and hotels that post staffing requests by shift and role — and an Admin console for the staffing companies, where hiring, shift assignment, and operations are managed. I joined as one of the first five employees, when the product already existed but was still rough: functional, but with an interface and an experience that were not where the business needed them to be.
The challenge
Three fronts at once
An early product to professionalize
The product already existed when I arrived, but its interface and experience were holding adoption back. It had to move from a functional-but-rough state to a quality bar that could carry commercial growth.
Three surfaces, three users, one ecosystem
JobSeekers, Clients, and Admin have very different goals and contexts of use. Each platform needed its own experience without breaking the coherence of the whole product.
Designing faster than the team could build
At an early-stage startup, design usually moves faster than engineering can implement it. The challenge was sequencing and prioritizing so every design decision turned into real product, not backlog.
Discovery
Research before the screens
Every platform started with discovery, not screens. For the first initiative — an app for the Client side — I ran a study of user profiles and needs through interviews and surveys. That same rigor carried over to the JobSeeker app: deeply understanding how a worker searches, filters, and applies for a job before defining a single flow. Research was not a separate stage. It is what ended up deciding what got designed, and in what order.
The project that unblocked growth
The JobSeeker app, from zero to production
I designed it end to end: research, user flows, the Figma component library, and every screen. Delivery was incremental — a validated screen moved to development while I pushed ahead with the next, and I worked alongside engineering component by component so the implementation never lost fidelity. When it shipped, the downloads, the applications, and the new clients followed on their own.
Decisions
Three decisions that defined the product
Betting on the side of the product with the most friction
The first app — for Clients — finished its design phase and moved into development, but an executive shift in priorities paused it before launch. Instead of a dead end, the research and learnings redirected focus to the higher-impact opportunity: a JobSeeker app that made job hunting radically simpler.
Design and build in parallel, without trading away quality
Rather than waiting for the full design before development started, I delivered the app screen by screen: one would get validated and move to engineering while I pushed ahead with the next. So speed would not cost quality, I worked alongside the developers building each component, making sure the implementation reached the same bar as the design.
AI-assisted request creation
I designed a job-request builder where the user could start from a text prompt, an Excel file, a photo, or voice dictation, and the AI assembled the full request: schedules, job types, and worker counts. One goal, four ways to reach it depending on how each client had already organized their information.
Featured feature
AI-assisted request creation
I designed a request builder where the client picks the starting point. The AI interprets any of four input formats and assembles the full request — schedules, job types, and worker counts.
✦ AI-generated
Structured request
Schedules, job types, and worker counts — ready to review and publish.
✦ Live component
Try the multi-input builder in the Lab
The pattern behind this feature is live and interactive: voice, text, and structured data converging on a single result. Try it yourself.
Execution
The component system
I built the component library that carried the app and worked side by side with engineering so every implemented component kept the fidelity of the design. The system became the quality reference for the rest of the product.
Results
The impact of the work
0 → production
JobSeeker app launched from scratch
5K–20K
Active users on the platform
100+
Staffing companies using Wolf
NY → Austin
Growth funded the new headquarters
Beyond the product
Brand, marketing, and conferences
Beyond the product, I designed pieces for marketing and customer service — material to promote the company and educate users — led two redesigns of the corporate website, and produced all the design for conferences: presentations, booths, and event material.
Takeaways
What I take away
- 01
Delivering design screen by screen, working with engineering on each component, keeps design quality alive in production. Speed does not have to cost fidelity.
- 02
When a project is paused by a business decision, the research is not thrown away: it is the input that makes the next bet sharper.
- 03
Modernizing one part of the product and leaving the other behind opens a gap that eventually gets paid for in conversion. Consistent quality across surfaces is a business decision, not an aesthetic one.
- 04
Being one of the first on a team means the designer does not just design: they set the quality standard everything else is built to.
On to the next project?