Personas are one of the core methods in a UX practitioner's toolkit. Traditionally, they have been created using qualitative research such as user interviews. While these approaches offer rich context, they often fall short when it comes to scalability and statistical grounding. More recently, generative AI tools have been used to produce personas, but studies have shown these outputs often reflect stereotypes and are not based on real user behavior. This talk brings science into the persona creation process by introducing a novel deep learning architecture that generates personas directly from customer data.
The result is a scalable, and mathematically grounded approach that outperforms machine learning methods and gives UX designers, researchers, and product managers a more rigorous way to understand and segment their users.
- Attendees will leave with a new perspective on how deep learning can enhance UX research
- a clearer understanding of the limitations of current persona practices,
- and actionable ideas for integrating large-scale data into persona development workflows.
About the presenter
Fabrizio (Fab), is currently working as a data scientist. Before that, he worked as a quantitative UX researcher, applying algorithms and statistical models to understand user behavior at scale and inform product design. Fab earned his Ph.D. in Human-Computer Interaction (HCI) from Rice University, where his research focused on symbolic AI and developing algorithms to scale and automate usability studies. He genuinely enjoys what he does, and spends much of his free time studying math or coding for fun. When he’s not doing any of those, you can usually find him dancing salsa or shredding on the guitar.
![[Talk] Beyond Sticky Notes: Deep Learning Meets UX Persona kapak fotoğrafı](https://secure.meetupstatic.com/photos/event/4/f/2/e/event_528080270.webp "[Talk] Beyond Sticky Notes: Deep Learning Meets UX Persona")
