About Me

I've worked as an undergraduate research assistant during my college, working in theoretical condensed matter physics and computational material science. I'm interested in materials and molecular discovery and in building AI4Science.

I have worked in first principles calculations - primarily in Density Functional Theory, where I used Quantum Espresso and Phonopy to study the thermal, vibrational and phononic properties of Graphene, Silicenes and other 2D nanosheets of group 14 elements.

My current research focus is on Thermal Rectification- achieving unidirectional heat flow in special 2D nanomaterials by tuning the substrate strain and performing first-principle simulations to confirm low-fidelity systems where the above phenomena can be observed- leading to the development of Thermal Logic.

I have also worked on Graph Neural Networks (using deep learning frameworks - pytorch and the Deep Graph Library (DGL))to predict molecular properties, in which I used GNNs to predict HOMO-LUMO gaps of small organic molecules (from the QM9 dataset).

I've previously worked in Machine Learning, specifically in Natural Language Processing where I built and developed a BERT and Spacy based question generation system for exam prep. 

Currently I'm working on a project- ASI- Artificial Scientific Intelligence- an agentic AI copilot for automating scientific workflows. Being a researcher, I'm familiar with the pain points scientists face, especially in the field of molecular discovery and The vision behind ASI is hugely motivated by that. My goal is to cut down on 1000s+ of hours of repetitive work that hinder throughput and automate the scientific pipelines using AI agents by integrating multiple tools like Quantum espresso, VASP, LAMMPS etc.