Research Fellowship

Offer Description

The research activity focuses on developing an adaptive self-supervised continual learning framework for speaker verification on hardware-constrained devices, such as smart eyewears. The goal is to design robust pseudo-labeling strategies and memory-efficient contrastive learning mechanisms to improve verification boundaries dynamically. By processing streaming unlabeled data fully on-device, the project aims to overcome ambient noise domain shifts and data scarcity while ensuring efficiency and user privacy.

Where to apply

Website: https://aunicalogin.polimi.it/aunicalogin/getservizio.xml?id_servizio=1079

Additional Information

Eligibility criteria

• Oral test aimed at ascertaining candidates’ aptitude and suitability to carry out the research activity covered by the Fellowship, as well as at assessing their knowledge of English and/or other languages relevant to the research activities to be performed (up to 30 points)
• Relevance and pertinence of the publications, theses and scientific products attached to the research programme covered by the Fellowship (up to 30 points)
• Relevance and pertinence of previous research activities and work experience, if any, in relation to the research activity covered by the Fellowship (up to 30 points)
• Relevance and pertinence of their study programme to the research programme covered by the Fellowship (up to 10 points)

Eligible destination country/ies for fellows: EUROPE

Eligibility of fellows: country/ies of residence: All

Eligibility of fellows: nationality/ies: All

Selection process: In order to participate in the selection, please read the call ("bando") available at the following website: https://www.polimi.it/en/bandi-incarichidiricerca

Website for additional job details: https://www.polimi.it/en/bandi-incarichidiricerca

Work Location(s)

Number of offers available: 1
Company/Institute: Politecnico di Milano
Country: Italy

Contact

City: Milan
Website: https://www.polimi.it/