2018 Intake Profiles

• Matteo Ceschia

  In my perilous journey through spacetime, I first graduated in Economics and Finance (BSc) at Bocconi University in Milan. But I could not resist physics sirens' songs, so I enrolled in the MPhys Physics at the University of Southampton. I was lucky enough to transfer to the particle physics flagship programme, so I ended up at CERN in my final year. My thesis focused on optimising the clustering algorithms used in the ECAL L1 trigger, with the goal of improving the identification of electrons and photons. I decided to join UCL CDT in Data Intensive Science because I think it is the ideal place to improve my data analysis skills, having the opportunity of applying ML models to both science and industry projects. I am now part of the SuperNEMO experiment, which aims to observe neutrinoless double-beta decay for the first time in the (not so) brief history of humankind. This would help us understand the fundamental properties of the neutrino and why there is an imbalance between matter and antimatter in our universe.

• Ishan Khurana

  I completed my masters at UCL researching multivariate techniques to improve the Higgs to bb analysis using monte carlo simulations of ATLAS data. The project included looking at using adversarial neural networks to ignore the effects of systematic uncertainties originating from monte carlo simulation.

• Augustin Marignier

  I graduated from UCL with an MSci Geophysics with First Class Honours. For my final year project, I performed a fully non-linear inversion of a large dataset of Rayleigh wave ellipticity measurements to create regional models of crustal seismic wave velocities using a high-performance computing cluster. I also developed and implemented a new model parameterisation based on splines for more physically realistic models. I took particular interest in the computing and data analysis part of this project, which led me towards the CDT in data intensive science. For me the CDT is a chance to further develop my computational skills and investigate the data analysis techniques used in different disciplines such as seismology and cosmology.

• Mario Morvan

  I graduated with both a Dipl. Ing. from Mines Saint-Etienne School and a MSci in Astrophysics from the University of Barcelona. While in the former I specialized in Data Science and Big Data, I later had the opportunity to apply these skills to astronomical data during my research project at the UB, developing a new method to detect open clusters from Gaia measurements. Seduced then by the combination of Machine Learning and Astronomy which, I realized, would continue to flourish in the near future due to the upcoming massive astronomical surveys, this CDT seemed to me like the perfect next step, and I therefore joined it after 8 months working in a Data Consultancy company in Paris.

• Sunil Mucesh

  I recently graduated with a Physics with Astrophysics (MPhys) degree from the University of Leicester. In the fourth-year, I worked on a project titled "Dissecting the Milky Way's disk(s)". The main aim was to determine if the Milky Way's disk can be separated into two distinct components. However, instead ofusing traditional analysis, I applied unsupervised machine learning algorithms such as self-organizing map. At the end of the project, I got the opportunity to present my research at the National Astronomy Meeting which inspired me to pursue a PhD. In particular, I wanted to apply machine learning in my research and as a result I applied for the CDT in Data Intensive Science.

• Constantina Nicolaou

  I graduated from Imperial College London with a BSc in Physics with First Class Honours. For my final year project, I analyzed Supernovae Type Ia (SNIa) data within a Bayesian Hierarchical framework to infer the cosmological parameters governing our Universe which fueled my passion for cosmology and motivation to embark on the journey of a PhD. Upon finishing my BSc degree, I was awarded the Imperial College UROP award for an 8-week summer research placement in the Astrophysics department. During my summer placement I continued developing my work on SNIa and got the chance to apply machine learning techniques to extract the brightness, colour and stretch parameters of a simulated SNIa dataset. Furthermore, I had the opportunity to use the Imperial College HPC cluster and parallelize code effectively. My experience during my summer placement inspired me to delve deeper into the world of machine learning and data intensive science which is the reason I applied to the CDT in DIS. Following my summer placement, I embarked on an MSc in Astrophysics at UCL. I graduated with distinction and was awarded the Harrie Massey Prize for best overall MSc student. In my master's project, I worked on the calculation of the Hubble constant from gravitational waves and in particular investigated statistical uncertainties and their impact on the inferred value of the Hubble constant. This was a very intriguing project which I really enjoyed and hence I am delighted to continue working on gravitational waves in my PhD project.

• Jeremy Ocampo

  I graduated with an MPhys degree from Oxford. In my final year project I was working with horn antennas which were designed for detection of weak THz signals from space, the goal was to optimize the power coupling between a receiver/transmitter antenna with respect to the profile of the horn antenna, this was achieved using software which utilised a genetic algorithm. I also have a keen interest in deep learning/computer vision/image processing in which I've taken extra courses during my degree. I've also taken part in machine learning competitions for fun. My main programming languages are Python, C++, and Matlab. I chose this CDT because of the range of data science skills you can learn/apply in the courses, and the wide variety of projects that the CDT has to offer.

• Katya Richards

  I graduated with an MMath from Cambridge with a focus on statistics and joined the UCL DIS CDT and Culham Centre for Fusion Energy, where I plan to use the power of machine learning to predict disruptions before they happen and save the world.

• Sam Van Stroud

  I have an MSci in Natural Sciences, specialising in particle physics, from Durham University. During my master's project, I applied deep learning techniques to classification of LHC collision data. Given the 4-momenta and ID of particles created in collisions, we want to be able to distinguish intersecting 'signal' events, such as those involving Higgs bosons with their uninteresting 'background'. Multivariate techniques such as neural are ideally suited to these. I have experience in bioinformatics and cybersecurity from internships with LabGenius and BT. At UCL, I will be working with Tim Scanlon to improving b-tagging efficiencies at ATLAS.