← Index

Case Study 04

Raylu

Founder sketches into investor demos.

Role
Sole designer, working with cofounders
Timeline
April 2023 to October 2023
Tools
Figma, React, Notion
Tags
AI Tooling · Web · Founder Engagement · Early-Stage
Raylu cover

Context

Raylu is an early-stage company building UI for machine learning and AI implementation tooling. The cofounders hired me to translate sketches and verbal product descriptions into a working visual identity, a Figma component set, and clickable prototypes for fundraising and early-adopter conversations.

Engagement

Working solo with two cofounders, I owned the design end of the product from sketches to functioning Figma prototypes. The work covered:

  • Brand exploration and identity
  • Competitive and analogue analysis
  • Translation of cofounder sketches into low-fi and hi-fi mockups
  • Clickable prototypes for investor demos
  • Pipeline-builder UI (the core product surface)
  • Component patterns (nav, modals, form structure) sized for React handoff

Process

Setting the Table. At the onset of the engagement I built a Figma workspace stakeholders could navigate without designer intervention. A linear timeline housed brand elements, competitor analysis, and product feature developments in vertically expanding sections.

Concepts to Prototypes. Cofounder sketches were translated through Zoom-based exploration sessions into concrete goals for look, feel, and functionality. Iterative low-fi to hi-fi loops kept founders close to design decisions without becoming a bottleneck.

Pipeline UI. The core surface was a workflow builder for AI/ML model pipelines. Initial nav structure split the app across Pipelines, Deployments, and Analytics. The detail screens cover pipeline creation (Q&A builder, templates, blank, duplicate), data file connection, environment configuration, and scheduling.

What Shipped

  • Brand identity and visual direction
  • Working Figma prototype demonstrating end-to-end pipeline creation flow
  • Component patterns (nav, modals, form structure) that mapped 1:1 to the React implementation and carried forward as the engineering team's working library

Reflections

The Figma workspace as alignment artifact was the highest-leverage decision in this engagement. Sketches and Zoom calls produce ideas. They don't produce shared context across founders, potential investors, and an external designer. The workspace became the document of record, which let founders self-serve pitch updates and kept me from being the bottleneck for every external conversation.

Pre-PMF prototyping has a different success metric than shipped product. The brief was "evidence the product is real enough to fund and build," not "polished system." Scoping down polish at this stage protects founder burn rate and lets the product hypothesis lead. The patterns left behind for the React team were the parts that needed to scale; the rest was sketch-rate by design.

Notes

Engagement focused on pre-product-market-fit prototyping for fundraising. Polish was deliberately scoped down in favor of clarity of concept.

Visuals