As a doctoral student and postdoc in chemistry and physics, I conducted computational research in a wide range of topics, from laser/matter interactions to nano-particle catalysis and kinetic modeling of bio-fuel reactions. While this research provided me with the necessary mathematical and programming experience to obtain an industrial data science position, it was only in my recent work developing machine learning algorithms that I recognized the potential for data science in the analysis of the relat ively small data sets that were available in my work as a physical scientist.

After a brief introduction to my previous work, I will discuss some of the challenges and solutions in my current work performing traditionally cloud-based ML with limited processing power and memory. Using my previous work as an example, I will discuss how pattern recognition and variance quantification can inform the analysis and interpretations of this data.