Research Group
Soňa works at KInIT as a Research Engineer, focusing mainly on audio tasks, including speech to text benchmarking, text normalization, transcription post-processing, model fine-tuning, and inference quality improvements. She also contributed to the SensAI project, where she generated LLM-based synthetic datasets published on Hugging Face and used for downstream fine-tuning. In addition, she develops predictive machine learning and deep learning models for customers, from architecture selection and model preparation to performance evaluation.
She holds a degree in Operational Research and Econometrics from the Faculty of Economic Informatics. Before joining KInIT, she focused on developing predictive models to forecast pricing trends and assess business performance, with particular attention to model explainability, data insights, and building tools for assessing machine learning models. She also contributed to policy evaluation research at the Slovak Academy of Sciences, where she analysed labour market data to assess the impact of active labour market programs.
Selected activities
She actively supports initiatives that inspire young girls to explore education and career opportunities in IT. Over the past three years, she has participated in Aj Ty v IT Girls’ Day. She also delivered a talk at “From Theory to Practice: Ready for the Future” event, sharing her career journey in IT and encouraging students to consider IT as a possible future path.
