Katarina Vukosavljević
DPhil Student
Before starting my DPhil, I completed an MEng in Engineering Science here at the University of Oxford. I have also gained experience in other areas as Automatic Speech Recognition, Computer Vision, and Software Engineering, through internships in academia and industry, but had become most curious about applying AI to healthcare. In particular, for my bachelor thesis, I worked with Prof Tim Denison and Prof Harrison Steel, and my team designed a non-invasive mouthguard for the detection and treatment of Obstructive Sleep Apnea. Then, for my master thesis, I worked with Prof Tim Denison, Prof Tingting Zhu, and Oxford Martin School's AGENDA team on designing a preliminary screening tool for epilepsy. I focused on developing a neural network for IED detection to aid the epilepsy diagnostic process in low- and middle-income countries.
As an EPSRC-funded CDT student, I was able to explore different areas of Health Data Science in my first year. For my first project, I worked with Prof Andrew Zisserman, VGG group, on using Masked Autoencoders for spinal radiological grading of MRI images. For my second project, I worked with Prof Tingting Zhu, χ lab, on modelling static and dynamic variables for outcome prediction using hospital data. I am mostly interested in researching and learning more about using deep learning methods for early diagnostics.
This year, I have moved to the Institute of Biomedical Engineering and joined the Computational Health Informatics or χ lab as a DPhil student. I will be working with Prof Tingting Zhu and Prof Julia Hippisley-Cox on designing a risk prediction tool for cancer. We will be using observational primary care data and utilise machine learning / deep learning methods.
When I am not working, you will probably find me doing fun things with my friends. I also like to spend my free time running, painting, or playing the flute!