My interests in pattern recognition ignited during maths Olympiads in my early education in the Sofia High School of Mathematics in Bulgaria. I came to the UK for my BSc in Mathematics with Management from the University of Leicester. Thrilled about machine learning, just before it had become so popular, I went to do my PhD at Aston University where I worked on scalable inference for Bayesian nonparametric models. In 2016, I joined the R&D team at ARM Cambridge to work on developing novel embedded hardware devices for occupancy estimation in smart buildings. Following my work on sensor monitoring, I continued with a cross-institutional postdoc funded by UCB Pharma to work on novel tools for the health monitoring of Parkinson's disease. In 2019 I joined the mathematics department at Aston University to help with the Rolls Royce Digital Academy in Data Science. In the fall of 2021, I started as a Horizon Assistant Professor in Data science to expand my work on statistical machine learning for various problems in the digital economy and healthcare.
My background is in statistics and computer science. The unifying theme of my research is interpretable machine learning in complex manifolds. Most of my theoretical work has been developing novel statistical machine learning algorithms for clustering, dimensionality reduction, feature sharing, and sequence modeling. My applied work has mostly focussed on algorithms for digital sensor monitoring applications and precision medicine.
Based on my own experience, I have developed a very inclusive teaching philosophy aimed at encouraging students from a variety of backgrounds and different learning strengths. My goal is to stimulate… read more
My current work is focused on robust estimation and scalable inference of domain informed clustering, dimensionality reduction, and segmentation problems in the context of precision medicine, digital… read more
Based on my own experience, I have developed a very inclusive teaching philosophy aimed at encouraging students from a variety of backgrounds and different learning strengths. My goal is to stimulate student engagement by provoking examples and games, with varying focus and complexity. For introductory and earlier level courses in more applied subjects, I advocate a constructive alignment philosophy. For higher-level courses, I am in support of flipped learning components which allow students to form their own opinion and expertise about the subject.
My current work is focused on robust estimation and scalable inference of domain informed clustering, dimensionality reduction, and segmentation problems in the context of precision medicine, digital healthcare, protein engineering, and automated aircraft engine maintenance. I am also looking for ambitious and enthusiastic PhD candidates interest in developing expertise in theoretical and applied machine learning or wearable monitoring.