cv

PhD in Computational engineering

TU Berlin, Berlin, Germany

M.Sc in Simulation Sciences

RWTH Aachen University, Aachen, Germany

B.Tech in Mechanical Engineering

National Institute of Technology, Durgapur, India

Research Associate & PhD Student

TU Berlin

• Developed and analyzed nonlinear Model Order Reduction (MOR) frameworks for complex reactive flows, including wildland fire propagation and published results in peer-reviewed journals.
• Formulated reduced-order optimal control problems for fluid dynamics applications by deploying adjoint-based optimization.
• Constructed custom numerical solvers in Python and C++, applying Finite Difference Methods (FDM) to efficiently solve complex PDEs for transport-dominated flows.
• Constructed custom neural network based optimization algorithm in collaboration with researchers from Aix-Marseille university for highly nonlinear and nonconvex optimization goals.
• Analyzed Rotating Detonation Combustion Engine (RDCE) test data by deploying ADMM equipped with nonlinear MOR to successfully separate transport phenomena and decouple highly nonlinear detonation wave interactions.

Undergraduate Research Assistant

TME, RWTH Aachen

• Coded hydrodynamic bearing calculation routines in FORTRAN for multi-body simulation software (MSC Adams / Virtual Dynamics).
• Developed a Thermohydrodynamic (THD) model by implementing the Reynolds lubrication and energy equations to accurately calculate local oil film temperatures and kinematics.
• Executed Finite Element simulations to analyze turbocharger thrust bearing designs, utilizing both mass-conserving (cavitation) and non-mass-conserving approaches.
• Accelerated computational speed and software robustness by integrating sparse solver techniques into the simulation architecture.
• Validated simulation accuracy by successfully correlating computed hydrodynamic pressures and forces against physical friction measurements from the test bench.

Machine Learning Intern

Launchpad.ai

• Engineered a promotional recommendation model utilizing XGBoost, GBRT, Python, and PyTorch, optimizing performance through rigorous hyperparameter tuning on synthetic datasets.
• Developed a novel clustering algorithm to automatically extract hidden cohort structures, statistically proving their impact on maximizing prediction accuracy for user engagement.

Google Summer of Code (GSoC) Intern

MBDyn

• Architected a bi-directional software coupling to integrate the open-source MBDyn kinematic solver with a pre-compiled Unsteady Vortex Lattice Method (UVLM) library.
• Engineered synchronized, real-time data transfer between subsystems to seamlessly exchange kinematic data and aerodynamic forces at every time step.