Research

Here, I learned about new modulation techniques for single-molecule break junctions, enabling precise control and measurement of molecular electronic properties. By varying contacts, anchoring groups, and bias, I learned how these factors influence orbital alignment and quantum conductance, providing deeper insight into charge transport mechanisms at the molecular and nano scale. As an undergraduate student, my chemistry knowledge was not deep enough to understand the intricacies of nanoelectronics. Therefore, my role mainly consisted of data collection using Scanning Tunneling Microscope Break-Junction (STM-BJ) and applying machine learning models for analyzing stochastic conductance data. My specific work has not been published yet in a paper; however, linked below is a paper involving similar work from my lab instructor, Bingqian Xu. 

I developed a novel Scientific Machine Learning framework for high-frequency electromagnetic modeling that embeds Universal Differential Equations within classical transmission line theory to achieve unprecedented accuracy in RF circuit simulation. Every component—from neural network design and simulation-based data generation to training and evaluation—was handled individually by me, supported by mentorship from Dr. Prathamesh Dinesh Joshi, Dr. Raj Dandekar, and Dr. Sreedath Panat. My approach integrates physics-informed neural networks within the Telegrapher’s equations, successfully capturing skin effect corrections and complex frequency-dependent phenomena that classical models miss. I pursued this project to support my goal of an M.S. in RF and wireless communication. This work is currently pending publication. The full research paper is shown below.