Pinki Khatun

Dr. Pinki Khatun

Postdoctoral Researcher

Member of the Collaborative Research Centre SFB 1294 – Data Assimilation

Institute for Mathematics

University of Potsdam

Email: pinki.khatun@uni-potsdam.de |pinkikhatun.nla@gmail.com

Research Interest

Numerical Linear Algebra
Preconditioning
Scientific Machine Learning

Biography

Dr. Pinki Khatun is currently a Postdoctoral Researcher at the University of Potsdam, Germany, where she is a member of SFB 1294 – Data Assimilation, working under the supervision of Prof. Melina Freitag. Her current research focuses on the reconstruction and prediction of the near-Earth radiation environment, with particular interests in model order reduction, data assimilation, and numerical linear algebra.

Previously, she worked as a Postdoctoral Researcher in the Department of Industrial Engineering at the University of Florence, Italy, under the supervision of Prof. Margherita Porcelli, on the project “Numerical Methods for Optimization Problems on Manifolds,” within the framework of MOLE (Manifold Constrained Optimization and LEarning). Her research at the University of Florence focused on numerical optimization on manifolds, randomized numerical linear algebra, and efficient computational methods for large-scale optimization problems.

Before joining the University of Florence, she worked as a Postdoctoral Fellow in the Department of Mathematics at the Indian Institute of Technology Gandhinagar (IITGN) under the supervision of Dr. Abhinav Jha, from May to October 2025.

She pursued her Ph.D. in Mathematics at the Indian Institute of Technology Indore (IIT Indore) under the guidance of Prof. Sk. Safique Ahmad, and defended her doctoral thesis in May 2025. She holds an M.Sc. and B.Sc. (Hons.) in Mathematics from Aliah University, Kolkata.

Dr. Khatun's research background is in Numerical Linear Algebra, with a particular focus on developing efficient and scalable iterative methods and preconditioners for large-scale linear systems and saddle point problems. Her research contributions include the development of robust preconditioners for Krylov subspace methods such as GMRES, as well as the derivation of structured condition numbers and backward error analysis for stability and sensitivity studies.Her research background is in Numerical Linear Algebra, particularly the development and analysis of iterative methods and preconditioners for large-scale linear systems and saddle point problems. Her contributions include preconditioning techniques for Krylov subspace methods such as GMRES, along with structured condition numbers and backward error analysis for studying the stability and sensitivity of numerical solutions. Her broader research interests encompass randomized numerical linear algebra, optimization, and scientific machine learning, with an emphasis on mathematically grounded and computationally efficient methods for scientific and engineering applications.

Page updated

Google Sites

Report abuse