Schnapps: AMAZON Aurora Operational Console
Role: Design & Front-end Lead
I led the complete redesign and future vision of the product, conducting extensive research through interviews with both existing and prospective users. I developed the user interface using the Cloudscape design system, collaborated closely with back-end developers, and built the entire stack from Python Flask and native JavaScript. This included creating a declarative UI from scratch and front-end development using a React application.
Product: Schnapps is an operational console for Amazon Aurora, a global-scale relational database service built for the cloud, fully compatible with MySQL and PostgreSQL. Schnapps enables Aurora operators to manage their service metadata and perform key control operations like deployments and configuration updates from a single interface.
The challenge was to transform an internal operation console, which was piecemeal and had been contributed to by hundreds of engineers over the past decade, into a structured, public-facing console. I was the sole designer, working alongside a team of 13 back-end and full-stack developers. I personally coded the application, committed the code, and built the architecture for an intuitive UI, bringing a disciplined approach to the teams across the organization to adhere to the design patterns.
Actionable Items from User Interviews
Personalization Features: Implement a customizable user profile where users can configure and prioritize frequently used features or tools.
​
Improved Visual Interface: Redesign the UI to enhance clarity and usability, making the interface more intuitive and efficient for users navigating complex operations.
​
Enhanced Result Accuracy: Improve the reliability and clarity of system statuses such as:
In-progress, Not refreshed, Success
​
Information Prioritization: Reorganize and prioritize critical information on the platform to enable quicker access—especially for on-call system engineers who need to retrieve data rapidly under time constraints.
Header Optimization: Re-evaluate header components:Reduce emphasis on the Search Box, which is rarely used.Maintain or enhance visibility of Cluster Selector and Tiny URL, which are among the most frequently accessed items.
Research Process
I conducted thorough research and user interviews to understand how the product was used, focusing on the most utilized pages and features. The goal was to simplify the interface by reducing the number of pages and features, as many tasks were being duplicated across multiple features. By segregating teams based on leadership within the organization and scheduling targeted interviews, I identified usage patterns and frequently used features. Interestingly, many teams used individual features 90% of the time primarily for retrieving or downloading metadata. This insight revealed an overlap of features, though less extensive than anticipated.
​
Further discussions with the engineering team revealed that they had custom formulas and documents to run programs for specific information retrieval. This led to an increase in tables, as each new team onboarding meant a new table to facilitate their specific needs quickly. While we couldn't stop teams from building pages for their convenience, we introduced an 'open page' with a saveable query feature, functioning like an open search to streamline operations.
​
This approach aimed to enhance the developer experience by making the platform more intuitive and less cluttered, ultimately improving efficiency and user satisfaction.
I conducted a card sorting activity with both current team members and former contributors who had worked on the core product. The goal was to understand key factors—particularly around the navigation system. This exercise helped me logically group items in the side navigation and gain valuable insights into how and why the system and UI were originally structured.
Since I was the first and only designer working on the product, and the entire platform had previously been built and maintained solely by backend engineers, it was crucial to first understand their thought process. This allowed me to refine the structure and begin introducing design thinking into the platform in a way that respected and aligned with the existing technical foundation.
