I am a Research Engineer in the Research Engineering team at Utrecht University.

I work on using generative AI for performing large-scale research data processing, in a wide range of fields. Together with domain researchers, I build pipelines for data processing, focusing on scientific rigour, engineering robustness, and computational reproducibility. 

I use generative AI for performing a variety of tasks that utilise the ability of generative AI models to interpret natural language, as well as their inherent general knowledge. I have used this technology for performing natural language processing tasks like information extraction, text classification, and sentiment analysis.  

I have used generative AI models from proprietary model providers - the GPT series of models from OpenAI, the Claude series of models from Anthropic, and the Gemini series of models from Google - as well as open-weights models like GPT-OSS-120B accessed through SURF's AI-Hub/WiLLMa.

These are some projects in which I applied generative AI to perform large-scale data processing:

1. Seedlists (Botanical Gardens): Extraction of plant species names from historical seedlists (catalogues of seeds) published by botanical gardens. 
2. Health Technology Assessment (HTA) documents (Pharmaceutical Sciences): Extraction of certain data points (name of drug, name of health indication, relative effectiveness, cost effectiveness, etc.) from documents published by HTA organisations in the EU.  
3. Kickstarter (Economic Geography): Assignment of industry codes to projects on the crowdfunding website Kickstarter. 
4. Historical Dutch newspaper articles (History): Assessing sentiment (positive/negative/neutral) in Dutch newspaper articles about fossil fuels, from 1960s-1990s. 
5. Scoping review (Human Geography): Performing a scoping review of young people's geographies. 
6. Biographies (Historical Sociology): Extracting sociodemographic variables from biographies from the Oxford Dictionary of National Biographies. 

If you are a researcher with a data processing problem that you think can be solved using generative AI, get in touch with the Research Engineering team by sending an email to research.engineering@uu.nl, or via this form.