Research Project
Development of a high-fidelity multiphysics computational framework integrating computational fluid dynamics, electrostatics, and particle tracing to investigate charged aerosol transport and optimize electrostatic particle collection by charge spherical collector.
| Duration | 2023–2024 |
| Role | Ph.D Researcher |
| Affiliation | Department of Mechanical Engineering, IIT Delhi |
| Methods | Computational Fluid Dynamics (CFD) • Finite Element Method (FEM) • Electrostatics • Lagrangian Particle Tracking • Numerical Simulation |
| Software | COMSOL Multiphysics • Python |
Developed an interpretable machine learning framework integrating LightGBM, SHAP, and Partial Dependence Plots (PDP) to analyze charged particle rebound, identify dominant physical mechanisms, and optimize electrostatic particle collection.
| Duration | 2024–2025 |
| Role | Ph.D. Researcher |
| Affiliation | Indian Institute of Technology Delhi |
| Methods | Gradient Boosting (LightGBM) • Explainable Artificial Intelligence (SHAP, PDP) • Feature Importance Analysis • Surrogate Modeling • Computational Fluid Dynamics |
| Software | Python • LightGBM • SHAP • Scikit-learn • COMSOL Multiphysics |
Developed a clustering-guided surrogate-assisted optimization framework integrating machine learning, K-means clustering, NSGA-II, TOPSIS, and explainable AI to optimize electrostatic particle collection while minimizing particle rebound.
| Duration | 2024-2025 |
| Role | Ph.D. Researcher |
| Affiliation | Indian Institute of Technology Delhi |
| Methods | Surrogate Modeling • K-means Clustering • NSGA-II • TOPSIS • Explainable AI • Multi-objective Optimization |
| Software | Python • LightGBM • Scikit-learn • pymoo • SHAP • Pandas • NumPy |
Solving Bayesian Inverse Problems using Variational Autoencoders for Uncertainty Quantification
OngoingDeveloping Bayesian deep learning frameworks for inverse problems by integrating variational inference, deep generative models, Fisher information geometry, and uncertainty quantification for computational science and engineering.
| Duration | 2026–Present |
| Role | Researcher |
| Affiliation | Indian Institute of Technology Pallakad |
| Methods | Bayesian Inference • Variational Autoencoders • Markov Chain Monte Carlo • Fisher Information Matrix • Scientific Machine Learning |
| Software | Python • TensorFlow • PyTorch • NumPy • SciPy • Matplotlib |