UC Berkeley
Data Analytics Software Engineer
Research and developed time-series classifiers and underlying feature algorithms. Applied these classifiers and algorithms to astronomical datasets with sample selection biases... more which arise from different instrument, scheduling, and survey characteristics. Developed crowd-sourcing and active-learning web applications which are then used to bootstrap existing classifiers onto new datasets. After several iterations of developing a classification pipeline and evaluating the effectiveness of the resulting classifier, the classifier is then applied to either real-time data streams or static survey datasets for scientific discovery of interesting or anomalous sources.
As the primary developer at Berkeley's Center for Time Domain Informatics, I collaborated with statistics and astronomy researchers to develop classification projects, one being Berkeley's real-time "Transients Classification Pipeline". This project incorporated machine learning to identify and classify science from the PTF telescope's nightly data stream.
As the primary developer at Berkeley's Center for Time Domain Informatics, I collaborated with statistics and astronomy researchers to develop classification projects, one being Berkeley's real-time "Transients Classification Pipeline". This project incorporated machine learning to identify and classify science from the PTF telescope's nightly data stream.
