Yongyi Zang

The core of my research centers on helping people use powerful AI. How capable a technology is and whether people can make good use of it are two separate questions (for example, computers were already powerful before the graphical interface, but it was only with windows, icons, and a mouse that most people could actually use one). In the context of AI, I’m interested in the second question: model capability is growing quickly, while people’s ability to understand and steer these models has not kept pace.

My long-term goal is to build a set of tools and methods that allow people to understand, steer, and verify the work done by increasingly capable models, rather than simply hoping that a model completes the task end-to-end.

I currently serve as Director of Research and Labs at Smule, where I lead a team working on models for audio and visual signals and work with the product team to surface these techniques to users in intuitive ways. In my independent research, I focus on interfaces for working with agents—organizing an agent’s work and where it might go next into a form people can grasp quickly, so that users can make the decisions that matter without reading through all of its output.

Before this, my background was in music production. These days I sometimes help friends write songs, and occasionally record covers of tracks I love.