Pulse

Pulse is a startup building an AI powered data room for founders raising money. The product has two sides: a founder side for sharing documents and managing an investor pipeline, and an investor side where investors review shared materials and run diligence with Perry, Pulse's AI assistant.

Before designing anything new, I needed to understand what already existed. Using Claude Code, I audited the codebase, mapping every component, every token (or lack of one), and every place the same UI pattern had been implemented differently. The audit gave us a shared picture of the work ahead: no design system, no tokens with roles, no documented components, no naming conventions.

Pulse didn't need rebuilding. It needed a foundation, and a structural, intuitive navigation and layout.

This is how I worked, with structure over surface. 🏃

I joined Pulse's sprint cadence and ran the work in weekly design cycles, anchored in the audit. Each week I iterated rapidly with Claude Design, built proposals in Figma, and brought them to Christian and Ipalibo.

I built it in Figma against the live codebase, with names matching at every layer: variables, components, React props. The MCP kept design and code in sync, so the system stayed live with the product, not parallel to it.

Step 01

Joined the repo branch and used the Claude Code MCP plugin to bring the current designs into my Figma file.

Step 02

Rebuilt the token layer with semantic roles for color, size, and typography, all meeting WCAG AA.

Step 03

Documented and fixed the core components, with naming consistent from Figma to code.

The design system became the source of truth for everything that followed.

Founder side.

The founder side is where founders share documents and manage their investor pipeline: Drive for files, Contacts for investors, Links for shared rooms. But it didn't feel like a fundraising tool. The interface wasn't intuitive enough, and that friction eroded the trust founders need in the tool running their raise.

To rebuild that trust, I had to redesign around what founders actually do during a raise, not the features Pulse already shipped. I started with a Jobs to be Done audit, a framework that asks what a user is trying to accomplish in each session, then anchors every design decision in that intent.

These five jobs became the lens for every layout decision. The redesign worked at three levels.

Investor side.

The investor side had the same architecture problem as the founder side. When reviewing user feedback, we found that new investors felt uncertain when opening a data room. The app used to open with an empty screen, “Select a Data Room.” Perry, which was supposed to be the differentiator, was hidden in a right side panel, only noticeable when you clicked on a file. And the diligence checklist disappeared the moment you opened a document.

I used the same framework and UX principles to approach this side, and the answer was simpler than the founder's. An investor opens a data room to do one thing: read the materials and make sense of them.

Those questions reframed the entire layout around a single principle: Perry is the differentiator, and the interface should prioritize it as such. From there, the redesign came down to four moves.

The landing page needed to feel as solid and professional as the product itself, a clear showcase of Pulse's features and visual identity.

I started from the PRD and translated it into a low fidelity wireframe to align with stakeholders on structure, hierarchy, and copy before moving into the final, high fidelity design.

Pulse needed me to build the design system from scratch, on a sprint pace. I started by documenting every component carefully, then reevaluated my priorities: design first, documentation later. That experience taught me how valuable my own workflow systems are: when they're ready from day one, time goes into design instead of organization.