Andrew Zahrt

For much of its history, organic chemistry has been a science driven by experimentation and intuition. However, since the turn of the 21st century, the field has experienced a renaissance for computer- and data- driven workflows for reaction development. One challenge that inhibits routine application of these methods in some areas of chemistry is the disconnect between how much data is routinely accessible to most optimization campaigns and the types of data-driven workflows available. Our research program seeks to resolve this disconnect through: (1) developing new workflows requiring minimal amounts of data, (2) reducing the experimental burden of data-driven campaigns with automated experimentation, and (3) using machine learning to pioneer new domains of optimization. The data-driven strategies herein complement existing workflows, enabling practitioners to select the right tool for the task at hand. Further, these tools are designed to be flexible—although we implement automated equipment and computational chemistry pipelines to maximize efficiency, they are designed to be accessible to chemistry laboratories without specialized equipment or expertise. Using this technology, we seek to expedite reaction optimization, catalyst design, reaction analysis, and reaction discovery, developing useful synthetic reactions along the way. We are also interested in the application of computational methods as a means of mechanistic inquiry.