Blaize AI Studio

Bringing sharper focus to AI modeling

A user-guided interface that focuses on the problem space rather than the tools.

As more industries move to employ AI systems, testing and iteration of these technologies becomes ever more important. This can be a laborious, murky process involving command-line code. Blaize’s AI Studio solves this problem elegantly, through an in-depth, user-guided experience crafted by argodesign.

AI Studio makes it easy to test and deploy models on chip architecture, eliminating the complexities of integrating disparate tools and workflows. It is a unique innovation in modeling that helps the user understand the tradeoffs between performance and precision.

An image of Blaize data prediction configuration

Turning design into code—and back again

The task at hand requires a unique interface and therefore was a greater challenge to build. Rather than page-based interactions, everything in AI Studio takes place in one view that the user never navigates away from.

To ensure this singular interface met the team’s design expectations, argo led an engineering team to make the front end a reality, turning an untraditional interface into working code that could then be deployed by Blaize’s back-end engineering team. A comprehensive Design Language System (DLS) using React Javascript and Storybook ensured that the built product matched the designs.

Two Blaize screens showing product workflow

Fidelity at the deepest level when needed

The resulting UI is made up of many ‘magic moments’—simple yet informative guideposts explaining what the user is doing so they can continue to focus on the problem they’re trying to solve. These guideposts, displayed in easy to understand natural language, help more novice users overcome common sticking points like locating the correct files, datasets, and parameter types.

The UI was designed with speed-to-deployment and iteration in mind. The top level UI elements display the simplest, but most complete path. This allows a user to quickly get something up and running, then go back through the process as needed to fine tune. Every card object in the interface can be clicked into to uncover a more sophisticated detail view for that point in the process. Here, deep configuration can happen. But novice users don't have to be bogged down with this more nuanced sophistication.

An image of blaize deployment configuration

Potential applications for AI Studio span multiple industries. Any instance that demands large-scale pattern recognition can benefit: self-driving cars, threat detection in security situations, medical imaging, package inspection, and more.

Creating a code-free platform for AI requires an understanding of the complexity users face, the unique nature of how AI is surfaced in the user interface, and how to simplify it all without getting in their way. argo has rare expertise among design firms in this area and was able to dig into the crux of the problem much faster. The resulting interface lifts AI Studio out of its highly technical underpinnings to make it usable for almost anyone.

Dmitry Zakharchenko

VP Product Management, Blaize

Next Project


Augmented intelligence for all