May 2026-present: A new evaluation tool EvalASR
A fast python package to evaluate ASR predictions. Besides the WER/SER/CER metrics, includes evaluation on entities and OOV words (recal/precision/F1), SemanticDistant, MER/WIL/WIP, and multiple visualisations
2021-2026: Ph.D. at the University of Zurich, Switzerland
Thesis Contextualisation of Automatic Speech Recognition and Related Applications
2022-2025: Call Center ASR (collaboration with Uniphore), IDIAP, Switzerland
Contributed to production-oriented ASR systems for call center applications, focusing on scalability, domain adaptation, and robustness to real-world variability
2020-2022: Air-Traffic ASR (collaboration with DLR and Fraport), IDIAP, Switzerland
Developed streaming ASR systems for air-traffic communication achieving WER below 5% in noisy, safety-critical environments
2016-2020: MA in Computational Linguistics and Speech Technologies at the University of Zurich, Switzerland
Thesis: Acoustic modelling for Swiss German ASR
2019: Sentiment inference and stance detection (participating as a student assistant), University of Zurich, Switzerland
2012-2016: Corpus of spontaneous Russian as a part of (Corpus of standard Russian), Saint-Petersburg, Russia
2014-2016: Role of pitch in the preservation of unity of “broken” discourse units in spontaneous speech, Saint-Petersburg, Russia
2012-2014: Recognition of ambiguous fragments of speech signal (evidence from spontaneous Russian), Saint-Petersburg, Russia
2007-2013: MA in Phonetics and Theoretical and Applied Linguistics at the University of Saint-Petersburg, Russia
Thesis: Sound contractions in Russian spontaneous and prepared speech