Suchi Saria


Suchi Saria is a Professor of Machine Learning and Healthcare at Johns Hopkins University, where she uses big data to improve patient outcomes. She is a World Economic Forum Young Global Leader.

Early life and education

Saria is from Darjeeling. She earned her Bachelor's degree at Mount Holyoke College. She was awarded a full scholarship from Microsoft. In 2004 she joined Stanford University as a Rambus Corporation Fellow. She earned her Master of Science and Doctor of Philosophy degrees at Stanford University, supervised by Daphne Koller and advised by Anna Asher Penn and Sebastian Thrun. At Stanford University, Saria developed a model that could predict preemie outcomes with a 90% accuracy. The model used data from monitors, birth weight and length of time spent in the womb to predict whether a preemie would develop an illness. The output value PhyiScore could be used to reduce the $26 billion per year spent by US health care on preterm birth. She worked in the startup Aster Data Systems.

Career and research

Saria believes that big data can be used to personalise healthcare. She is considered an expert in computational statistics and their applications to the real world. She uses Bayesian and probabilistic modelling. In 2014 Saria was funded by a $1.5 million Gordon and Betty Moore Foundation project that looked to make intensive care units safer. The project used data collected at patients' bedsides along with noninvasive 3D sensors that monitor care in patient's hospital rooms. The sensors collect information on steps that might have been missed by doctors; like washing hands.
Saria uses big data to manage chronic diseases. She is part of a National Science Foundation award that looks at scleroderma. She uses machine learning to analyse medical records and identify similar patterns of disease progression. The system works out which treatments have been effectively used for various symptoms to aid doctors in choosing treatment plans for specific patients. She has developed another algorithm that can be used to predict and treat Septic shock. The algorithm used 16,000 items of patient health records and generates a targeted real-time warning score. She collaborated with David N. Hager to use the algorithm in clinics, and it was correct 86% of the time. Saria modified the algorithm to avoid missing high risk patients- for example, those who have suffered from septic shock previously and who have sought successful treatment. She was described by XRDS magazine as being a Pioneer in transforming healthcare. In 2016 Saria spoke at about using machine learning for medicine at TEDxBoston. The talk has been viewed over 50,000 times.

Awards and honours

Her awards and honors include: