Glib possesses PhD in Physics. The main topics of his research were nonlinear dynamics, chaos and super-symmetric quantum mechanics. Those research topics require heavy use of numerical simulations, and his team had adopted CUDA early on after it’s arrival.
After leaving academia in 2014, Glib spent two and a half years working on a project, which aimed to find biomedical indications in human locomotion data. Based on data collected from wrist-worn accelerometers the team was able to determine various important indications and project successfully transitioned to product phase.
Shift to machine learning was natural, since both numerical simulations of dynamic chaos and data from various sensors feature the same structure: it’s time series. Currently Glib works on several projects, including predictive maintenance system for manufacturing, which analyzes thousands of sensors in near real-time to provide equipment operators with actionable insights on how to prevent failures.