New Faculty Q&A: Steve Mussmann

This fall, Steve Mussmann joined the School of Computer Science (SCS) as an assistant professor.

With a focus on data-centric aspects of machine learning (ML) systems, Mussmann develops methods and best practices to make ML systems easier to build, monitor, and improve.

“By studying data-centric ML, I hope to address critical bottlenecks for the broad adoption and use of ML, a promising tool to help address humanity's biggest modern-day challenges,” he said.

Before joining Georgia Tech, Mussmann was a full-time ML researcher at Coactive AI in San Jose, CA. He had previously been a postdoctoral fellow at the University of Washington after receiving his Ph.D. in computer science from Stanford University.

What interests you about working at Georgia Tech?

I was drawn to Georgia Tech because of its energetic, innovative, can-do attitude that drives impactful research and its supportive, open, and welcoming culture.

What is your research focus?

Motivated by the difficulties of building and deploying useful machine learning systems, I study data-centric aspects of machine learning, which are often overshadowed in the ML research community by the development of learning algorithms and machine learning models. More specifically, I study methods and best practices for data collection and selection, using alternative forms of human supervision (beyond gold labels) and ML system evaluation/monitoring.

How did you get interested in this field of research?

When I decided to join Georgia Tech, I had research experience in two different data-centric areas: active learning for data collection (the main topic of my Ph.D.) and data selection (the area I explored during my postdoc). In my year at Coactive AI, I had the opportunity to learn directly from and work with customers, hoping to leverage ML to solve their business use cases. Through this experience, I realized that the most widely used ML formulations insufficiently capture the full breadth of the challenge of creating a useful ML system. I was motivated to research data-centric ML to address many of these gaps.

What are you most looking forward to in your new position?

I'm most looking forward to learning and experiencing the breadth of interests and ideas in a vibrant academic community, mentoring and working with students, and researching solutions to important issues.