Bridging Data Science & Design
11:00 - 11:30 | Friday, May 24
We’re all developing new types of user experiences, ones heavily backed by machine learning. These data products present new design challenges and require that we evolve our approach to product development. Does the user understand our algorithmically generated content? Does she trust our recommendations? User experience is a critical part of the success of an algorithm. Reciprocally, the collection of data signaling successful user behaviors are key to the training of an algorithm. Yet, many data scientists and machine learning engineers work in silo from designers with minimal interactions if any.
In my talk, I’ll share practical ways to bridge design and data science with learnings from working on search at Airbnb. I’ll share approaches to
- Translation: how to bridge the differences in language and values
- Testing: rethinking what to experiment on and when
- ML strategy: how to design in service of your model with examples from work images, text and recommender systems.
Data Science Manager, Airbnb
Claire Lebarz leads one of the data science teams at Airbnb, with focus areas in search, personalization and causal inference. She manages initiatives ranging from analytics to machine learning applications. Prior to Airbnb, she led the data science team at Turo, a car-sharing two sided platform, building their pricing and marketplace intelligence. Claire is an applied mathematician and economist by training, she researched the link between economic inequality and instability, between the Paris School of Economics, the IMF and UC Berkeley. Claire is also an advisory board member of the Data Institute, partnering with the University of San Francisco.