Liminal
Liminal turns fragmented data into real-time, actionable intelligence through Link, an expert-led, AI-powered platform. I’ve led the product design for Link, shaping core features, creating the Link Mosaic design system, and improving collaboration across Product and Engineering through process design and hands-on workshops.
Context: Liminal
Year: 2024 — Present
Scope: Product Design, Design System, UX Research
Tools: Figma, Mixpanel, Hotjar
Allowing users to Catch-Up
As part of the Monitor feature in Link, I designed the Catch-up experience — a dynamic, animated flow that surfaces the 10 latest events from a user’s monitors. The goal was to help users quickly review what matters most while laying the groundwork for personalized insights. I crafted an interface that encourages users to rate each event, feeding into a tailored algorithm that will power future recommendations. This solution balanced clarity, engagement, and long-term value, while pushing the boundaries of how users interact with evolving data.
A new way of filtering
I redesigned the filters bar in the Monitor feature to move beyond static dropdowns and into a more contextual, task-oriented experience. Filters are now presented inline and adapt to the data shown, allowing users to not only focus on specific event types or timeframes but also actively exclude information that isn’t relevant. This dual filtering approach gives users greater control over how they engage with the data, making the experience faster, more intuitive, and tailored to their needs.
Tracking feature success
To support and validate design decisions in Monitor, I led the implementation of Mixpanel tracking with a UX research mindset. I defined key user interactions to track — such as filter usage, event ratings, and engagement with Catch-up — mapping them to core user journeys. This allowed us to measure feature adoption, identify drop-off points, and spot emerging usage patterns. The insights gathered informed improvements and helped align design, product, and engineering around real user behavior.
You've reached the end!
