Retrosynthesis is the process of recursively simplifying a desired target molecule into smaller and less complex starting materials. Retrosynthesis is a cornerstone for organic chemistry, offering pathways or synthetic routes to create important chemicals such as medicines and novel materials. Previous template-based approaches have used ML to predict the precursors given a target compound. This work demonstrates the addition of pathway level information improves the accuracy of the retrosynthesis model.

Through this SuperUROP, I want to use computational approaches to tackle chemistry problems. I wish to apply my classroom knowledge from chemistry and engineering to discover pathways to synthesize a target compound. I am excited about tackling chemistry problems from a computational angle, and I hope to improve my presentation skills and publish a paper after completing this program.