Building a product from scratch at an early-stage startup

Starting from zero

When I joined, there was no brand, no website, no design system, and no product UI. The backend was being worked on, but everything the user would eventually see and interact with didn't exist yet. The team was small and fully remote across the UK, France, USA, and Canada. The target audience was still being defined. Priorities shifted constantly between investor needs, pilot programmes, and building out the core product. There was no playbook.

Finding the foundation

The first job was giving the company a face. Before any product work could begin, there needed to be something to point people to - something that forced early decisions about tone, positioning, and who we were actually talking to. So I started with the brand: logo, identity, and the public-facing website, designed and built from scratch.

From there I moved into the product itself - an app where users create projects, conduct behavioural interviews, and generate reports. No existing patterns to build on. No design debt to navigate around. Just a blank canvas and a backend being built in parallel.

Building the system

With the product taking shape, I needed a design system that could scale beyond just me and that any developer could pick up without a steep learning curve. The approach was deliberately lightweight - building on established frameworks rather than creating everything from scratch:

• shadcn for components
• Lucide for icons
• Tailwind for styling variables
• Storybook for documentation

Familiar tools mean faster collaboration and fewer things breaking. Sprint planning in Linear kept design and engineering in sync, and I worked directly alongside the developer - using Claude Code and IDE tooling to close the gap between design and build.

Wearing every hat

With a team this small, the role stretched well beyond design. Brand identity, marketing website, design system, product UI, Storybook documentation, and stakeholder presentations all running in parallel, often shifting priority week to week.

But it wasn't just design work. I worked directly in the codebase alongside the engineers, committing via Git, reviewing pull requests, and using Claude Code to keep output precise and consistent. Storybook wasn't something I handed off to developers to populate. I built and maintained it myself, so components were documented, tested in isolation, and genuinely ready to pick up without a separate briefing. The gap between design and engineering was as small as I could make it.

What came out of it

The startup went from no design presence to a fully branded product: live website, working app, and a scalable design system ready for any developer to pick up. The Figma library and Storybook setup means components are documented, consistent, and maintained. The groundwork is there for the product to grow.

Towards the end of the engagement the brand direction shifted. We had begun exploring a refresh, but startups move fast and direction can change overnight. I took the new direction and used it as the foundation to realign the app, a useful reminder that the most valuable thing you can do in an early-stage environment is build things solid enough to adapt quickly, whatever comes next.

This is the kind of work that teaches you how everything connects. Brand influences product. Product constrains system. System shapes what's possible in the next sprint. Owning it end to end, and learning a completely new domain in behavioural science along the way, is the part I'm most proud of.