Rich Caruana

Also known as Richard Caruana

Rich Caruana is a machine learning researcher whose work covers multitask learning, ensemble selection and interpretable models for healthcare. After a Carnegie Mellon PhD in 1997 with Tom Mitchell and Herb Simon, he joined Cornell's computer science department in 2001, presented ensemble selection at ICML in 2004, and co-chaired KDD in 2007. His 2015 KDD paper "Intelligible Models for HealthCare", written with Yin Lou, Johannes Gehrke, Paul Koch, Marc Sturm and Noémie Elhadad, received the 2026 SIGKDD Test of Time Award for Applied Data Science. In 2019, he co-authored the InterpretML paper, which describes an open-source package with the first implementation of the Explainable Boosting Machine. He is Chief Science Officer at Intelligible.

Now at: Intelligible
Badges: Bridge
Loyal

Timeline

1997

PhD at Carnegie Mellon

He earns a PhD at Carnegie Mellon University in 1997, working with Tom Mitchell and Herb Simon. His thesis on Multi-Task Learning helps create interest in a new subfield of machine learning called Transfer Learning.
2001–2004

Ensemble selection at Cornell

He joins Cornell's Department of Computer Science in the fall of 2001. In 2004, he presents ensemble selection at ICML with Alexandru Niculescu-Mizil, Geoff Crew and Alex Ksikes, a method that builds ensembles from libraries of thousands of models, and receives an NSF CAREER Award for his work on meta clustering.
2007

Co-chairing KDD 2007

In 2007, he co-chairs the KDD conference with Xindong Wu and wins a best paper award with Daria Sorokina.
2014

Do deep nets need depth?

In 2014, he and Lei Jimmy Ba publish "Do Deep Nets Really Need to be Deep?" at NIPS, showing that shallow feed-forward nets can learn the complex functions previously learned by deep nets. He also wins a best paper award in 2014 with Todd Kulesza, Saleema Amershi, Danyel Fisher and Denis Charles.
2015

Intelligible models for healthcare

In 2015, he and coauthors Yin Lou, Johannes Gehrke, Paul Koch, Marc Sturm and Noémie Elhadad publish "Intelligible Models for HealthCare: Predicting Pneumonia Risk and Hospital 30-Day Readmission" at KDD. The paper shows that generalized additive models with pairwise interactions (GA2Ms) can match the accuracy of black-box models such as neural networks and random forests on high-stakes healthcare prediction tasks, while staying interpretable enough for a clinician to inspect and correct.
2019–2022

The InterpretML package

As a principal researcher at Microsoft Research, he coauthors the InterpretML paper in 2019 with Harsha Nori, Samuel Jenkins and Paul Koch; the open-source Python package includes the first implementation of the Explainable Boosting Machine. In 2022, he is a visiting scientist in the Simons Institute's Summer Cluster on Interpretable Machine Learning.
2026

A KDD Test of Time Award

In 2026, he receives the SIGKDD Test of Time Award for Applied Data Science, together with Yin Lou, Johannes Gehrke, Paul Koch, Marc Sturm and Noémie Elhadad, for the 2015 healthcare paper.
As of October 2026

Chief Science Officer at Intelligible

He is Chief Science Officer at Intelligible, a company led by founder and CEO Ben Lengerich that builds AI that shows its work. He continues to work with the team on tools that make interpretable, trustworthy models accessible to teams without a research lab's time or expertise.