Thomas Thies Email & Phone Number

London, England, United Kingdom

Contributing to the research and development of a machine learning powered application which helps charities improve the effectiveness of their digital fundraising and engagement.• Led the development of a deep learning text classification pipeline which uses 32 DNNs to extract information from our clients' content (and deployed the PyTorch models on AWS).• Built a deep learning recommender system in TensorFlow to predict supporter interests in unseen topics, based on the topics they engaged with (achieving a top-3 accuracy of 93%).• Developed and deployed gradient-boosted decision trees to detect supporters at risk of ending their recurring donations (using time-series data).

Villigen Psi, Switzerland

• Peer-reviewed comparative study of brain scans from two SOTA X-ray microscopy beamlines, aiming to guide developments in X-ray microscopy of biological samples at present and future synchrotron beamlines. • Paper soon to be published in the Optics Express journal.

Lund, Sweden

• Developed an image processing pipeline in Python for users of the division's X-ray CT scanner to obtain 3D reconstructions of their samples (using novel phase contrast methods).

Lund, Sweden

• Investigated the structural and magnetic properties of thin film materials by analysing thousands of X-ray diffraction patterns in Python.• Poster presentation at MAX IV conference.

Msc, Data Science With Artificial Intelligence, Distinction

MSc project: Handling computationally expensive multiobjective engineering problems with Bayesian optimisation and genetic algorithms Core.

Bsc, Physics, First-Class Honours

Core modules: Statistical Physics • Scientific Computing • Mathematics • Medical Imaging • Electronics • Condensed Matter • Radiation &.