T127 Reflection 4
Fourth reflection for T127 course — gallery walk and capstone showcase.
The gallery walk was one of the more interesting showcases during my time at HGSE. Most classes rely on presentations or pitches, which limit feedback to a single moment. The gallery walk inverts that. Both the projects and the audience felt curated, and the audience worked in the LDIT space or on the TLL team, which made the feedback on my capstone unusually pointed.
Our Data Fluency Course team decided to weave a story: problems surfaced in the discovery phase informing the design decisions in the build stage. My piece focused on helping users reach their modules faster through scaffolding and checks for understanding. Questions inside the modules nudge users to enter the terms their school or department actually uses, and those terms feed the universal data dictionary.
The idea generated a lot of interest and questions. People wanted to know how the architecture works and what the backend looks like. The open-source project is calibrated to fit a standard iframe (or code block) within an LXP, though some manual wiring is needed to set it up.
Beyond the novelty of Generative UI, people were curious about how the dictionary works, and I had to explain the nudging and personalization loop a few times. One minute is an extremely tight window to walk through this. The other recurring question was cost, which my calculation puts at ~$300 for 50,000 unique UI generations. As I repeated throughout the presentation, the app is designed around learning designers themselves: every part of the codebase can be calibrated by them as they see fit. Human-in-the-loop is the philosophy that anchored the whole design.