Tharindu Jayasinghe Email & Phone Number

Santa Clara, Ca, Us

San Jose, California, United States

Berkeley, California, United States

• Managed a NASA funded scientific program to discover quiescent black holes and neutron stars in the Galactic neighborhood using large spectroscopic and photometric datasets.• Led a spectroscopic follow-up program and developed a python based data analysis pipeline to measure the radial velocities (RVs) of binary stars using data from the Lick Observatory. • Developed python packages to model the spectral energy distributions and fit observed time series photometry of binary stars… Show more • Managed a NASA funded scientific program to discover quiescent black holes and neutron stars in the Galactic neighborhood using large spectroscopic and photometric datasets.• Led a spectroscopic follow-up program and developed a python based data analysis pipeline to measure the radial velocities (RVs) of binary stars using data from the Lick Observatory. • Developed python packages to model the spectral energy distributions and fit observed time series photometry of binary stars using Markov Chain Monte Carlo sampling.• Built an XGBoost classifier to identify ~580,000 RV variable stars out of a catalog of ~34 million stars.• Enhanced a deep learning framework that interpolates synthetic stellar spectra, and increased the number of pixels in the output by ~27x, allowing for the compression of an ~18 GB dataset into a ~10 MB PyTorch model (factor of ~1800x compression).• Managed and mentored a team of 4 to conduct spectroscopic observations and data modeling for >400 binary stars resulting in the detailed characterization of the physical properties of these systems.• Software used: numpy, scikit-learn, pandas, PyTorch, TensorFlow, matplotlib, SQL/AQDL Show less

Columbus, Ohio, United States

• Led the analysis of time series photometric variability of stars for the All-Sky Automated Survey for Supernovae (ASAS-SN).• Developed a machine learning based pipeline using the scikit-learn random forest classifier to identify and classify variable stars from a dataset consisting of >10 billion individual brightness measurements for 60 million stars in the Milky Way.• Published the first all-sky catalog of homogeneously identified and classified bright variable stars, totaling… Show more • Led the analysis of time series photometric variability of stars for the All-Sky Automated Survey for Supernovae (ASAS-SN).• Developed a machine learning based pipeline using the scikit-learn random forest classifier to identify and classify variable stars from a dataset consisting of >10 billion individual brightness measurements for 60 million stars in the Milky Way.• Published the first all-sky catalog of homogeneously identified and classified bright variable stars, totaling ∼660,000 variables and including ∼220,000 new discoveries, an improvement of 50% over existing catalogs. • Led the ongoing citizen science project "Citizen ASAS-SN", which has collected >1.5 million classifications of time series data from over 5,600 citizen scientists worldwide.• Published 12 first author peer-reviewed papers describing the scientific output of this project, including the discovery of rare variable stars and new insights into the physics of binary stars.• Software used: numpy, pandas, matplotlib, scikit-learn, SQL/ADQL Show less

Columbus, Ohio, United States

Doctor Of Philosophy - Phd, Astronomy

Thesis title: The discovery and characterization of variable stars in the All-Sky Automated Survey for SuperNovae Advisors: Prof..

Master Of Science - Ms, Astronomy

Bachelor Of Science - Bs, Physics, 4.0