Experience

Research Experience


MAIL LAB, Carnegie Mellon University

Research Assistant Aug 2022 - Present

Project1/Master’s Thesis: Predicting Interactions in Heterodimeric Complexes using Multimodal Data Integration: Integrating GeomEtry-Aware Relational Graph Neural Network (GearNet)and Language Models in a Supervised Framework, with Self-supervised Insights.

The project aims to advance the understanding of protein/subunit interactions within heterodimeric complexes by leveraging multimodal data integration and state-of-the-art machine learning techniques. By predicting interactions between protein residues, the project seeks to understand the underlying biological mechanisms governing complex formation, which could have significant implications for drug discovery and therapeutic interventions. Co-author paper currently in preparation for submission to NeurIPS 2024. Report

Project 2: Peptide Property prediction of short-length peptides using GNNs and PeptideBERT. The purpose of the project was to develop and fine-tune PeptideBERT, a state-of-the-art protein language model based on Transformers,specifically designed for predicting important properties of peptides. PeptideBERT was trained using ProtBERT, a pre-trained transformer model specialized for protein sequences. Published a paper.

Project 3: The purpose of the second part of the project is to develop a robust framework for predicting properties of short-length peptides by leveraging advanced techniques from graph neural networks (GNNs) and protein language models. Specifically, the project utilizes the fine-tuned PeptideBERT and a Geometry-Aware Relational Graph Neural Network as encoders to capture both sequence and structural information of peptides.

Generative AI Innovation Incubator, Carnegie Mellon University

Aug 2023- Present

Application of encoder-only models for the design of inhibitor sequences, followed by their optimization through Adaptive Exploration in Multi-Objective Thompson Sampling-Guided Directed Evolution.

Description: Implementing Thompson Sampling-guided Directed Evolution (TS-DE) framework for multi-objective protein sequence optimization, integrating adaptive exploration techniques to balance and optimize conflicting objectives simultaneously. Additionally, incorporated tree-based neural bandits to further optimize exploration strategies and accelerate the discovery of high-performing protein sequences.

MAIL LAB, Carnegie Mellon University

Graduate Student Researcher

August 2022 - April 2023

Molecular Chemistry Lab, Vellore Institute of Technology

Undergraduate Student Researcher

November 2018 – August 2022

Cellular Biomechanics Lab, Vellore Institute of Technology

Undergraduate Project Assistant

June 2019 – October 2019

Molecular Biophysics Unit, Indian Institute of Science

Undergraduate Summer Research Assistant

June 2019 – October 2019

Please do reach out to me through email, if you have any questions about my research. I’d love to talk more!

Return to Main Page