Currently I am the Lead Data Scientist at Popertee Ltd in Dublin, Ireland, leading the research and development of a location data platform for pop-up retail and experiential marketing campaigns.
Prior to this I was a Natural Language Processing Specialist at award-winning health-tech startup babylon in London, mostly concerned with chatbots, information extraction and assorted applications of biomedical natural language processing.
I’m also an occasional contributor and editor at Wide Orbits, an online community dealing with ideas, issues and culture based in Dublin, Ireland where I’ve written about topics as diverse as computational creativity and Brazilian politics.
How it came to this
I completed my doctorate in the Computational Linguistics Group at Trinity College Dublin under the supervision of Dr. Carl Vogel in 2013 entitled: “Translation Effects in English Natural Language text”as part of the Centre for Next Generation Localisation (CNGL), an SFI-funded CSET research centre.
In my thesis I looked at questions related to stylometry and authorship in the context of translations, applying machine learning techniques to tease out signals in literature.
After a brief foray deploying and training commercial information extraction systems in the Irish Times, I worked as a postdoctoral researcher at the Centre for Applied Data Analytics Research (CeADAR), based at University College Dublin, where I applied my knowledge of natural language processing to problems in social media analysis and social entity resolution. I also worked on graph analytics and data visualisation projects at CeADAR.
Over the years I’ve worked on diverse topics in text analytics such as:
- Applications of text analytics and machine learning to literary translations.
- investigating the stylistic patterns of character in dramatic works.
- Textual analysis of social media and news.
- Training and deploying commercial NER systems.
- Developing a prototype text alignment tool for translations.
- Quality estimation for translation.
- Identifying errors in non-native speaker English.
- Authorship attribution of translations.
- Diachronic classification of Latin and English text.
- Computational creativity and generation of apt titles for news articles.
- Biomedical entity and relation extraction.
- Biomedical natural language generation.
- Geolocation systems for short-term retail
Deutsch wird ab und zu gesprochen. Português também. Beagánín Gaeilge freisin.