Dr James O’ Donoghue

Dr. James (Jim) O’ Donoghue is a Postdoctoral Researcher at the Nimbus Gateway. His work has spanned 13 years across academia and industry.

Initially focussed on data warehousing, OLAP (online analytical processing) and data mining techniques, Jim was an early adopter of deep learning and deep neural networks. His PhD research in Dublin City University (DCU) was funded by the European FP7 In-MINDD (Innovative Mid-Life Intervention for Dementia Deterrence) project.

His research aim was to take deep networks outside of their traditionally used realms of natural language and computer vision and apply them to health. As such, his research looked to abstract approaches to high-dimensional and sequential data, automate hyper-parameter optimisation (now called AutoML), develop techniques to explain and interpret hidden features and layers, and create tooling to do so.

After his PhD research, Jim spent a year in DCU on various post-doctoral projects. He developed an interactive survey and analysis web-tool for the National Anti-Bullying Centre (ABC) and researched and developed techniques for agricultural trade data analysis and price optimisation.

Subsequently, Jim went into industry. He joined an Atlantic Bridge and Enterprise Ireland funded Trinity research spin-out, where he contributed to developing a graph database in Prolog.

From here Jim went on to contribute to large projects and lead teams in various domains. He developed a stock-level optimisation and demand forecasting tool in the retail sector; created IoT ML models and deployments via embedded systems and MQTT in the Agri-Tech sector; led the development of a parallelised cloud-stream messaging and decoding system for binary data, and predictive models for IoT GPS devices in the Logistic and Supply chain sector; as well as lead model development and deployment in the health-tech sector. In addition to these AI and ML based projects, Jim developed and contributed to other more foundational data warehousing and engineering solutions along the way.