..

highlights

• August 2025: I presented our paper “Speech LLMs in Low-Resource Scenarios: Data Volume Requirements and the Impact of Pretraining on High-Resource Languages” as a poster at Interspeech 2025 in Rotterdam :) Check out the paper or poster.

• July 2025: We at the FBK SpeechTek lab release a HuggingFace collection of pre-trained multilingual speechLLM projectors that can be downloaded and used for further fine-tuning or decoding. More information here!.

• May 2025: My first first-author paper as a PhD student has been accepted at Interspeech 2025! In our paper titled “Speech LLMs in Low-Resource Scenarios: Data Volume Requirements and the Impact of Pretraining on High-Resource Languages”, we systematically investigate how much training data is needed for a SpeechLLM linear projector to be comparable to a Whisper-only setup for ASR. We then show that pretraining these projectors on high-resource languages can effectively mitigate performance issues typically associated with data scarcity. More details to be shared later on!

• November 2024: I started my PhD in Information Engineering and Computer Science at the University of Trento and Fondazione Bruno Kessler. I am supervised by Alessio Brutti and Marco Matassoni, and work within the SpeechTek lab.

• October 2023: I graduated with honours and obtained a M.Sc. in Cognitive Science from the the University of Trento, Italy. My thesis focused on investigating the feasibility of speaker-independent deep learning voice conversion approaches to improve speech from English speaking patients with dysarthria. I was supervised by Dr. Alessio Brutti and Dr. Gianluca Esposito

• August 2023: I attended the 12th ISCA Speech Synthesis Workshop in Grenoble, France. Preliminary results from my M.Sc. thesis were accepted as a late-breaking report and presented as a poster at the workshop.

• September 2022 - October 2023: I was a research intern at the Speech Technology Lab at the Bruno Kessler Foundation (FBK).