- Research on decoding techniques for single channel multi-talker speech recognition (cocktail party problem),- Research on noise robust automatic speech recognition for air-traffic management, including acoustic modelling and contextual language model adaptation,- Supervising projects, teaching numerical and computer exercises in several courses: Signals and Systems, Bayesian Models for Machine Learning and Machine Learning and Recognition.

- Development of automatic speech recognition system for the U.S. court reporting service,- Working with both hybrid ASR systems and E2E ASR system including the Transformers.

- Research on continuous source separation

- Design and development of light-weight ASR and object detection systems and their deployment into 5G antenna (BTS), i.e. a device with constrained computational resources,- Lightly-supervised training of Spanish ASR system based on subtitled speech and experiments with incremental lightly-supervised learning.

- Development of a mobile POS application for Android tablets.- Creation of automated tests for mobile apps.- Infrastructure management.

- Creation of automated tests.- Continuos deployment of middleware.