Lead designer contributing to several observability projects for an enterprise cloud data services dashboard used by thousands of businesses globally.
This is a reproduction of a 0-1 dashboard I worked to prioritize and streamline with 6 PMs, 1 UXD, and 9 Devs to scope out the MVP and deliverables within 3 weeks to present to leadership.
Google Dataproc
UX Design, Interaction Design, Visual Design,
AI Strategy
Define
Problem Statement:
How might we streamline a troubleshooting dashboard for data professionals to find and solve observability problems quickly?
Dataproc is a complex Google Cloud platform that data professionals use to monitor when systems are down.
As Dataproc grows, dashboards become more complex. Dashboards may include multiple filtering and hard to understand data viz.
As workloads increase, Dataproc customers’ overall costs may spike.
Feature Priotization
In a brainstorming session, I scoped initial MVP features of Dr. Proc with 1 UXD and 1PM to focus on a dashboard showing a central summary of troubleshooting. This solution would help our users minimize information access cost.
The dashboard would focus on our critical user journey showing: What happened, why this occurred, how we can fix it.
Implementation
Developed and implemented 0-1 creation of a MVP dashboard to manage serverless batches and integrate AI in troubleshooting.
MVP implemented
Example of A/B Test dashboard data viz exploration tested with real users
Testing & learnings
Reduce complexity. Users were overwhelmed by multiple types of data viz and did not understand distinctions.
Design for Engagement. Filtering by time frame was unclear and users did not engage with the double filter
Next steps (Iteration)
5 PMs were added to my pod after MVP and had multiple ideas to integrate Gemini AI.
I drove UX strategy towards focusing on 1 clear way to integrate AI, use of visual icons, and scoping the user journey into multiple steps.