I hold a PhD in Machine Learning from Imperial College London and a MEng in Electrical and Computer Engineering.
Currently, I'm a Senior Research Scientist at Kheiron Medical Ltd, working on Deep Learning algorithms to assist radiologies in breast cancer detection.
I have a total of 3.5 years of industrial experience in positions that allowed me to be at the forefront of recent developments in Deep Learning and Computer Vision.
Previously I had the opportunity to work on:
- Deep Learning for visual perception in autonomous driving at Toyota Motor Europe
- Virtual assistants and Bayesian insurance claims modelling at AIG
- Deep Learning for image understanding at Cortexica Vision Systems
Senior Research Scientist
Kheiron Medical Ltd, Machine Learning team
As a senior member of the Machine Learning team my main focus is identifying and pursuing innovative directions inspired from recent academic advances in the field. I also actively contribute to the improvement of the company's core products and the design of our shared codebase.
Senior Engineer | Deep Learning | Autonomous Driving
Toyota Motor Europe, R&D Department
As part of the Recognition Technology team at Toyota, I was responsible for developing Deep Learning methods for vision-based automated driving.
My efforts were focused on both adding new functionality as well as making the underlying algorithms more computationally affordable.
More specifically I have had the opportunity to work on state-of-the-art network architectures for Semantic Segmentation, Depth Estimation and Visual Odometry, using Caffe and Tensorfow.
Nov. 2016 - Oct. 2017
Machine Learning Scientist
AIG, Science Department
As part of the Science Department and the Special Projects Lab at AIG, I have had the opportunity to work on a few innovative projects within the business.
More specifically, I proposed a complete solution to build a virtual assistant at AIG (similar to Siri, Cortana or Facebook M), including appropriate methods and data collection process. I investigated models based predominantly on Recurrent Neural Networks for intent tracking and slot-filling in interactive dialogue systems.
As part of my work on this project, I also created a demo application in Python for the purpose of showcasing the potential benefits to the business.
I also worked on the following projects:
a) Real-time Object Detection using Tensorflow.
b) Risk modelling for insurance claims data, using hierarchical Bayesian methods.
May 2016 - Nov. 2016
Machine Learning Researcher
Cortexica Vision Systems, Research Team
Cortexica is a computer vision start-up, spin-off from Imperial College, providing content-based image retrieval to several major retailers (JohnLewis, Macy's)
For over a year, I was responsible for devising machine learning algorithms focused towards improving the retrieval pipeline.
Finally for almost a year I served as Scrum Master of the Research Team, responsible for applying Agile principles and practices.
Oct. 2014 - May. 2016
Imperial College London
During my PhD I had the opportunity to work with a very wide variety of predominantly Bayesian methods for the purpose of modelling human motion.
I worked extensively on Bayesian non-parametrics and Dirichlet processes, Gaussian processes, Gaussian process latent variable models (GPLVM) and many more. I was performing approximate inference by means of Variational Bayesian and MCMC sampling (Gibbs, HMC and ESS)
More specifically, during the second half of my PhD I worked on Bayesian matrix and tensor factorisation for the purpose of rectifying severely damaged time-series in an unsupervised manner. Similar approaches are frequently utilised in recommender systems.
Oct. 2010 - Dec. 2014