Cofounder & Chief Scientist, Brightband
I'm a versatile physical scientist who enjoys working with fast-moving teams on challenging and important problems.
I've worked across diverse domains, from cosmology to remote sensing, from mineral exploration to weather forecasting. The common thread: making sense of sensor data with physics, statistics, and machine learning.
In 2024 I jumped back into the world of AI weather forecasting as cofounder and Chief Scientist of Brightband. Previously I was a Staff Data Scientist at KoBold Metals and Chief Scientist at Descartes Labs. Before that I was a cosmologist at the University of Chicago and Stanford, working on the South Pole Telescope, a project that took me to Antarctica five times.
My 2022 solo paper showed that a machine learning model could forecast global weather on par with the operational physics-based models from NOAA and ECMWF at 1° resolution. Cited 450+ times, it helped spark today's wave of AI weather models.
paper · talk · Science · NYTimes · project site · code
Deep learning-based visual search over the Earth. A precursor to today's geospatial foundation models.
paper · blog
Petabyte-scale satellite image processing on commercial cloud, back in 2015: one petabyte in one day.
2015 paper · 2017 paper
One of my first deep learning projects, trained with Alex Krizhevsky's cuda‑convnet. A fun bridge between my work in cosmology and my future work in machine learning.
site
Highly cited measurements of the cosmic microwave background, including my Ph.D. measurement of the damping tail (450+ citations).
Ph.D. paper · B-modes · all publicationsOther papers from my cosmology days: CMB lensing · lensing by clusters · neutrinos & the CMB · cluster velocities · no “dark flow”