Viktor Gregor

Research areas: machine learning, statistics, speech processing, natural language processing, large language models

Position: Research Engineer

Email
LinkedIn

Viktor’s current research focuses on automatic speech recognition (ASR) for Slovak, with the broader goal of bringing Slovak speech technology closer to the level available for major world languages. His work spans rigorous evaluation methodology, model fine-tuning, and curation of speech data. A significant part of his work involves cooperation with industry partners, as well as developing KInIT’s own audio services.

With a background in applied mathematics, Viktor spent around 10 years as a data scientist in the gaming industry before joining KInIT, leading projects in predictive modelling, recommendation systems, causal analysis, Bayesian experimentation, and LLM-based analytics. Beyond research, he contributes through workshops, expert volunteering, and mentoring.

Selected Student Supervising

  • Pavol Lisý – Application of statistical models for prediction of player behaviour in mobile games – Master thesis supervisor, FMFI UK 2020
  • Martin Majer – Application of random forest for user activity prediction – Bachelor thesis consultant, FMFI UK 2018