Improving Data Annotation UX for ML Teams
Ivan Tszian
Overview
Designed a custom dashboarding system for an AI platform, allowing users to create flexible dashboards and visualize key data in one place.
Goal
Enable users to build custom dashboards and charts to monitor relevant metrics without relying on predefined views.
Challenges
The key challenge was designing a flexible system that supports different use cases without becoming overly complex.
Users needed powerful customization, but the interface had to remain intuitive and easy to use.
Solutions / Outcome
Designed a modular dashboard system with customizable charts and flexible layouts.Focused on clarity, usability, and scalability to support different data needs.
The feature was successfully shipped, giving users more control over how they explore and monitor data.
Overview of annotation queues with statuses, progress, and reviewers.
Trace view showing inputs, outputs, and annotation results.
Hover states on graphs
Annotator view
Creating evaluation fields for annotators.
Like this project
Posted Apr 11, 2026
Reducing friction in high-volume annotation and review workflows
