I am a computational chemist turned data scientist, working as a consultant at Gyant and expanding my data science/machine learning knowledge on the side. I also love baking sourdough bread and generally experimenting in the kitchen, as a way to use my chemistry degree in real life!
During my PhD at UC Berkeley, I worked on modeling the large-scale behavior of molecules attached to the surface of nanoparticles, known as ligands, and investigating the structures they formed. While doing so, I developed my skills coding in Python for analyzing molecular simulation data on the order of hundreds of gigabytes. I also added functionality to LAMMPS/PLUMED, open source softwares for molecular dynamics, to perform importance sampling in parallel and understand the probability distribution of ligand geometries and their physical properties. Since experimentalists are unable to obtain images of these ligands, this project provided a detailed insight into how these ligands behave that is otherwise not easy to obtain, and can be used to predict nanoparticle properties for developing technologies like superconductors, catalysts and solar cells.
Since graduating, I’ve been working as a data scientist at Gyant, classifying unlabelled user texts to improve product performance. I’ve learned how to use NLP tools and transfer learning for embeddings to be able to better classify data from the Gyant chatbot and point users to appropriate endpoints based on their queries. It’s been great fun to be part of a growing team of smart people who are building products to make our healthcare systems better.
I really enjoyed the aspect of using data to answer these difficult questions and predict trends that can be useful in deciding new directions and areas to focus on. I’m looking forward to doing so more and more on my journey as a data scientist!