The view from halfway: quantifying the star formation histories of massive galaxies at z ~ 1

High-quality spectroscopy of distant universe galaxies is limited and expensive to obtain, but reveals the stellar absorption features that directly trace star formation histories (SFHs) of a galaxy where as multi-wavelength photometric data is relatively abundant and contains information about total stellar mass and dust reddening within a galaxy. Together they help us decipher the processes that drive stellar mass assembly of galaxies. We leverage the Large Early Galaxies Astrophysics Census (LEGA-C) survey's high-quality spectroscopic data of ~3000 massive (log($\mathrm{M_*/M_{\odot}}$) > 10.5) galaxies in COSMOS field in a sweet spot of z ~ 1, where each galaxy was observed for ~20 hours. We characterize star formation histories of this mass and volume complete sample to shed light on two critical questions of galaxy evolution - 1. When and how rapidly did galaxies form their stars and did this depend on other fundamental properties like stellar mass or stellar velocity dispersion? Does this inference change using different modeling tools and priors? 2. How often do galaxies diverge from the traditional evolutionary path and restart star formation (‘rejuvenate')? How does this compare with UniverseMachine theoretical model predictions that incorporate dark matter assembly?. Both of these analyses required full Bayesian SPS modeling of LEGA-C galaxies using different prescriptions of SFHs (parametric - Bagpipes and non-parametric - Prospector). This research has been computationally expensive, requiring efficient parallelization and was supported by >5 million CPU hours of resources from the Pitt Center for Research Computing (Pitt CRC).