Neuron

Across three continents

The CognitiveScale design team was asked to build a variety of products without a shared visual language. The company had development teams in Austin, Hyderabad, and London, each with its own technology stack, making it difficult to build consistent experiences. In addition, CognitiveScale’s customer base was shifting to a new industry, finance, with different software expectations. It was time to create a design system to design a unified family of software, to unify tech stacks, and to establish a design language that spoke to our new customers.

Establishing a brand

After studying the finance industry and the software they used day to day, we decided to establish our design language as a confident and bold voice. Over the course of several moths we build components, grid systems, custom icons, color schemes, and then applied them to the products we worked on.

A vehicle for development

Well that previous paragraph was pretty hand wavy. In order to really build out a functional design system we needed a vehicle. A real product that could serve as a testing ground for all the components and patterns we wanted to try out. That product was Cortex Console, a DevOps resource management application. It offered opportunities to establish navigation patterns, form patterns, and visual language. I designed Cortex Console, feature by feature, and built the design system as I want. With this proving ground, the new design system was able to be tested in the real world. As the Cortex Console project finished up we settled on a name for the design system, Neuron.

Neuron Rollout

Here are two examples of how Neuron was applied to future products at CognitiveScale. The first of which was a product called…

Cortex Toolkit

Cortex Toolkit was a tool for annotating text for for Natural Language Processing. CognitiveScale’s Data Scientists used the Toolkit to prepare data in preparation for training Machine Learning Models. Neuron provided the backbone for this project and helped a fellow designer and myself move quickly to try out different iterations, testing them with our internal Data Science team. In addition the Neuron components, which had already been built helped the Hyderabad dev team implement our designs quickly.

Profile of One

Neuron even helped with products that had to be built largely from scratch, like the Profile of One visualizer. The visualization allowed people to zoom in and out to see the type of data built around a profile at different levels of granularity. Neuron provided myself and another designer on the team with brand elements, navigational structures, and contextual panels so that the product would feel like it was part of the larger Cortex family, despite requiring such extensive custom work.