Andrei Barcovschi

Biography & Research Interest

Andrei Barcovschi is originally from Chisinau, Moldova but has grown up in Ireland and completed his Bachelor of Engineering degree in Electronic and Computer Engineering at the National University of Ireland, Galway (previously NUIG, now University of Galway) in 2020. His bachelor thesis involved developing a custom neural Computer Vision pipeline for detecting, tracking and analysing traffic flow at a road intersection, deployable on a low resource embedded device. He further received a Master of Science degree in Artificial Intelligence at NUIG (University of Galway) in 2021, with his thesis focusing on researching efficient parallel neural vocoders for Text-to-Speech (TTS) synthesis on edge devices. He is currently a first year PhD student at the University of Galway, working in collaboration with industry partners on improving TTS and Automatic Speech Recognition (ASR) technologies for child speech understanding and synthesis. His focus is on researching neural models of different architectures and configurations to achieve these goals. His research interests include various machine learning techniques and tools, especially neural networks employed for solving a wide range of tasks in the domains of Natural Language Processing (NLP), including conversational AI, chatbots, TTS and ASR.

Research Activities

Working with State-Of-The-Art (SOTA) TTS and ASR models for child speech generation and recognition.

Working on large-scale synthesis of generated child voices for massive training data augmentation.

Collaborating in the development of informative evaluation studies on augmented speech data to help gauge quality of synthetic audio samples.