Conversational shopping through six specialized AI agents that understand style and learn preferences, deployed as a multi-agent architecture on GKE.
It is telling that the grand prize at the same event was also a shopping assistant built as coordinated agents. That says something about what this cohort of judges rewarded: not one model doing everything, but specialists with a visible division of labour. The number six matters less than being able to justify the split, and separating style understanding from preference learning is a defensible line.
Summaries are written by this site. Project pages include demo videos, screenshots and the team's own write-up (videos may autoplay).