Ipsita Ghosh

About

I am a PhD student in Computer Science at the University of Central Florida and a member of the UCF Institute for Artificial Intelligence. I work with Dr. Christian Kümmerle on optimization methods for machine learning, with interests in low-rank training, geometry reconstruction, and scalable non-convex optimization.

Before beginning my PhD, I worked as a Data Scientist at Micron Technology. I hold an M.Sc. in Data Science from Chennai Mathematical Institute, an M.Sc. in Mathematics from the University of Delhi, and a B.Sc. in Mathematics from St. Xavier's College.

Research

My research focuses on developing efficient and reliable optimization techniques for modern machine learning models. I am particularly interested in exploiting low-rank and sparse structure to reduce training cost while preserving model quality, as well as in geometry-aware learning problems that connect machine learning with numerical linear algebra and high-dimensional probability.

Recent work includes low-rank training methods for deep neural networks, sample-efficient geometry reconstruction from Euclidean distances, and optimization questions motivated by reinforcement learning from human feedback. I enjoy problems that combine mathematical theory with practical algorithm design.

Background

At Micron Technology, I developed data-driven methods for manufacturing systems using online change-point detection, Bayesian optimization, and self-supervised representation learning. I have also completed internships at Rutgers University (DIMACS), Haber, and Hopstack, working on optimization, deep learning, and sensor data analysis.

Recent Updates