Simulating life at the atomic scale.
University of Illinois Chicago — Advisor: Prof. Huan-Xiang Zhou • Expected graduation: July 2026
I use molecular dynamics simulations, enhanced sampling methods, and deep learning to understand how intrinsically disordered proteins interact with cellular membranes — with applications in neurodegeneration, signal transduction, and drug discovery.
Background
I am a fifth-year PhD candidate in Computational Biophysics at the University of Illinois Chicago, working in the lab of Prof. Huan-Xiang Zhou. My research sits at the intersection of atomistic simulation, machine learning, and bioinformatics, with a focus on intrinsically disordered proteins (IDPs) and their interactions with lipid membranes.
Prior to UIC, I completed a BS-MS dual degree in Physics from IISER Mohali, India, where I worked on liquid crystal biosensors and lipid-protein MD simulations (published in Nano Letters).
I have six peer-reviewed publications (four as first author), including papers in Cell Reports Physical Science, Communications Biology, and Annual Review of Biophysics. I also developed and deployed AroMIP, a public web server for predicting peptide-membrane insertion energetics with >90% accuracy on human-proteome disordered regions.
Microsecond-scale all-atom MD simulations of disordered proteins at lipid bilayers, uncovering aromatic insertion codes and aggregation pathways.
GNN-VAMPnet pipelines (PyTorch Geometric) for classifying membrane-binding states from MD trajectory data; automated conformational analysis.
Characterizing Aβ40/42 oligomerization and aggregation pathways using enhanced sampling methods, with cross-validation against NMR and cryo-EM data.
Sequence-based predictors deployed as public web servers — bridging high-throughput simulation with accessible tools for experimentalists worldwide.
Projects
My work combines atomistic simulation, large-scale sequence analysis, and machine learning to extract actionable biological insight from disordered protein systems.
Scanned 1.2 million 9-residue sequences via the PPM method to derive a sequence code for aromatic-centered membrane insertion. AroMIP achieves 91–99% accuracy on human proteome IDRs and is live as a public web server.
Microsecond AMBER/NAMD simulations with umbrella sampling and free energy calculations to map aggregation intermediates of Aβ40 and Aβ42. Identified key oligomeric states implicated in Alzheimer's neurotoxicity.
Engineered a hybrid Graph Neural Network + VAMPnet deep learning pipeline using PyTorch Geometric to classify disordered-protein membrane-binding conformational states directly from MD trajectory data.
Developed AMBER-compatible force fields for chemically challenging small molecules (free radicals, polyethylene glycol) using Gaussian-based RESP charge fitting, benchmarked against existing AMBER standards.
Led the computational arm of a collaboration with Mayo Clinic, integrating MD predictions with experimental NMR and cryo-EM to validate structural ensembles of amyloid-beta oligomers across multiple partner institutions.
Installable Python CLI + library that drives a full GROMACS MD workflow end-to-end from a single natural-language prompt via Claude Code. A 13-step DAG handles structure prep, equilibration, production, analysis, and reporting — with per-step fingerprinting and resume-after-crash. Ships with three Claude skills, 8 tutorials, and 148+ tests.
Performed all-atom MD simulations of 5CB liquid crystal self-assembly and lipid-protein interactions at IISER Mohali, supporting a Nano Letters publication that established LC droplets as nanoscale biosensors.
Peer-Reviewed Work
Six peer-reviewed publications (four as first author) · 87+ citations · h-index 5
Technical Expertise
Five years of hands-on experience across simulation, machine learning, and scientific software development.
Achievements & Contributions
Get in Touch
I am actively seeking industry and academic positions starting Summer 2026. Feel free to reach out to discuss research, collaborations, or opportunities.