Machine Learning Innovation Summit
November 29, 2018, Dublin
Senior Software Architect & Engineer
A Software Engineer/Developer, Technical Architect he leads a team of engineers for the machine learning platform team for preventing and detecting fraud in debit, retail and credit transactions. He specialises in building real-time/low latency distributed data and messaging systems using C#, C++ and Java. Intelligent real time machine learning applications are a game changer in any industry. In the transactional fraud detection industry, we have to make approve or decline decisions within 100 milliseconds. A lot happens in this short amount of time: data transformation, data tokenisation, feature extraction, machine learning model execution, behaviour modelling rules, decisioning rules and reporting. There is a common problem in moving machine learning from the data science environment into production. Common problems include rolling in and out model artefacts, data engineering (creating real-time/non-real-time data pipelines including data storage and security), feature engineering, model execution, rules execution, verification, continuously deployment, monitoring, analytics and more.