London, United Kingdom
London, United Kingdom
CCUK Credit Card Payment Defaulting Algorithms: Project Manager (04 months)
• Call Credit UK initiated a project to create algorithms to determine a credit card applicant’s probability of defaulting. Engaged as PM to lead this project. Analysed existing scoring systems; defined project approach; led 4-strong development team; determined relevant algorithms and data processing methods; wrote SQL queries and liaised with customer to ensure correct model calibration. Succeeded in delivering algorithms based on Random Forest, resulting in an increased accuracy of 30%, facilitating 20% reduction in losses.
RedZebra Analytics: Banking Customer Prediction Tool: Data Scientist (06 months)
• Red Zebra wanted to create a solution to predict customer behaviour and segmentation, using transactional data of a bank, which would allow them to match product offerings to customers. Engaged as Chief Data Scientist to create the solution. Analysed data models; identified transactional data patterns; created predictive models; utilised Self-Organized Maps, Neural Networks and Random Forests to created proof of concept. Succeeded in creating solution, allowing a new product line for RedZebra.
Vodafone Portugal: Customer Churn Prediction Solution: Data Scientist (02 months)
• Vodafone initiated implementation of a system to predict the probability of customers leaving their network. Engaged as Data Scientist to develop solution. Analysed complex inter-network data; liaised with stakeholders; worked with Apache Graphx to build graphs; created weighted graphs and developed stochastic diffusion model based on CDR data. Succeeded in delivering solution, resulting in Vodafone identifying potential departing customers & facilitating implementation of client retention processes and marketing strategies.
Turim Hotels, Portugal: Room Demand Algorithms: Data Scientist (02 months)
• Turim Hotels wanted to create algorithms to predict future room demand in order to determine relevant sales & pricing strategy. Engaged as Data Scientist to create algorithms. Analysed historical data; identified Google data trends; utilised Deep Learning neutral networks; scrutinised data from competitors; and separated data trends to account for seasonal variations. Succeeded in delivering algorithms, resulting in predicted sales accuracy of 90%, enabling identification of fallow periods which informed Turim’s sales & pricing strategy.
dataAI: Multimodal Image Recognition: Project Manager (07 months)
• dataAI requested an algorithm to do facial emotion recognition for clients, using images. Engaged as PM, created proof of concept; assigned developers; utilised Convolutional Neural Networks. Succeeded in creating solution - achieving human-level accuracy in identifying seven different emotional states within 30k image databases.
Inesting SA: Digital Marketing Dashboards: Project Manager (24 months)
• Inesting SA initiated a project (Digitaligence) to create a dashboard to aggregate data from different social media platforms & Google analytics; to monitor and determine marketing ROI. Engaged as PM to lead project. Identified key KPIs; led 3-strong development team; assisted creation of artificial intelligence model; and configured rules for monitoring activity. Succeeded in delivering dashboards, resulting in defined sales data that facilitated companies’ ability to create effective and targeted sales strategies.
Portuguese Science Foundation: RBS Algorithm: Data Researcher (2 years)
• Engaged as Principal Investigator in a scientific project using Artificial Neural Network algorithms to analyse Rutherford Backscattering (RBS) data. Analysed previous RBS experiments; scrutinised existing data; wrote code using C++; invented an algorithm called HLVQ; resulting in real-time analysis of complex RBS data & publication of paper in 10 international journals, including Journal of Applied Physics.
Portuguese Science Foundation: Credit Scoring Research: Project Manager (3 years)
• Engaged as Principal Investigator in a scientific project to create a credit-scoring algorithm to rate SME’s/Big companies (<1000 employees). Led team of 3 research students; developed project approach; designed dashboards and led implementation. Published 15 scientific papers outlining project findings. Succeeded in creating algorithm to calculate and track the credit scoring of 100.000 French companies.
Chief Data Scientist
January 2015 - June 2015
1997 - 2000