Romie Banerjee

Built prototypes for proof-of-concept, focusing on object detection, segmentation and video analytics machine learning pipeline for delivering products, implement state-of-the-art in deep learning for human detection, tracking, re-identification, build software library for modular AI functionality. Delivered GPU-based parallelized implementations. Researched scientific literature to propose solutions for open problems in AI and computer vision

Researched on topics in Category Theory, Algebraic Topology, Data Analysis, Topological Data Analysis, Machine Learning, Deep Learning, Compression/Quantization of Neural Networks, Knowledge transfer/distillation, Quantitative Finance, Stochastic Calculus. Taught undergraduate and graduate courses in various branches of Mathematics, Statistics, Computer Science, Machine Learning and supervised research.

Research in Algebraic Topology, Topological Data Analysis (TDA), Data structures encoding high-dimensional topological features of point cloud data, high-dimesional feature extraction using TDA and persistent homology, Machine Learning, Deep Learning, applications of TDA to Data Visualization and understanding Neural Networks. Presented research results in workshops and conferences. Invited talks at American Mathematical Society Joint Meetings Boston, Wayne State University and University of Regensburg

{ Taught undergraduate courses in Linear Algebra and Advanced Calculus. { Invited talks at University of Chicago and Brown University

Computational Algebraic Topology

Mathematics

Mathematics and Computer Science