FBK is a private research institution based in Trento (Italy), operating in different scientific fields and disciplines. As such, it keeps the Autonomous Province of Trento within the mainstream of international research. FBK comprises 12 research centers with activities and production available at http://www.fbk.eu/research-centers
Workplace
The Augmented Intelligence Center is a newly established FBK center aiming to bring together a plurality of scientific areas already present in FBK to open up a new frontier of research in the field of Augmented Intelligence. The key research lines in Augmented Intelligence encompass: Foundational Artificial Intelligence; Human-Artificial Intelligence Interaction; Cooperative and Social Artificial Intelligence; Proactive ethical Artificial Intelligence.
The FBK Research Center in Augmented Intelligence encompasses FBK research groups working in machine learning, natural language processing, dialogue technologies, speech recognition, human-computer interaction, computer vision, multi-modal perception, neuro-symbolic techniques, data and knowledge representation and reasoning, data and process intelligence, neuroinformatics, cooperative and social Artificial Intelligence. The Center benefits from the expertise and scientific excellence of approximately 80 researchers and developers, including senior researchers, emerging talents, and doctoral students affiliated with these FBK research groups.
The SpeechTek Lab at Fondazione Bruno Kessler investigates AI-based solutions for speech technologies, addressing a variety of speech related tasks: automatic speech recognition (ASR), speech enhancement and separation, spoken language understanding, speaker identification, automatic language learning and other speech and audio applications. The group is particularly interested in the application of AI topical research directions to the speech context, such as: continual learning, large-scale models, self-supervised adaptation, edge processing. These are key features towards an effective and efficient deployment of speech technologies in real-life conditions and are being investigated also in multi-modal (audio-visual) scenarios. Finally, the group has also a long-lasting experience in customization of ASR services.
FBK actively seeks diversity and inclusion in the workplace and is also committed to promoting gender equality.
To promote the inclusion of disabled staff as per law 68/99, the Foundation is available and interested in evaluating the applications received for technical-scientific domains that do not correspond exactly to this call.
Job Description
The SpeechTek Lab is currently involved in two Horizon Europe focusing on evolving language technologies for EU languages: Eloquence (eloquenceai.eu) and Meetween (meetween.eu).
We are looking for a researcher who will help us investigate, develop and deploy novel solutions for multilingual speech recognition and spoken language understanding, leveraging pre-trained foundation models (also multimodal) as well as large language models. Particularly relevant are:
- low resourced settings, typical of many European languages, which requires efficient training or fine-tuning in terms of data;
- compliance with European values and regulations in terms of bias, ethics and privacy.
While working full time on the two projects, the candidate is expected to advance the state-of-the-art in the field, disseminate the results to top international conferences and journals of the speech processing and artificial intelligence communities.
Job requirements
The ideal candidate should have:
- Master degree in data science, computer science, engineering or related fields (the degree must be obtained by the start date of the FBK contract);
- Fluent in Python and in the ML packages and modules (pytorch, tensorflow);
- Experience in AI solutions for speech recognition and speech processing;
- Excellent publication record in relevant international conferences and journals;
- Good communication and relational skills;
- Self-motivation and result orientation;
- Oral and written proficiency in English.
Additional requirements:
- PhD degree or 3 years of equivalent experience in fields related to this call;
- Familiarity with most common pre-trained models and neural architectures for speech and audio processing (WavLM, Wav2Vec,...)
Employment:
Type of contract: fixed-term contract
Working hours: full-time (38 h per week)
Gross annual salary: about € 36.200 - € 41.400 depending on background and expertise in the field.
Start date: preferably by December 2024 / January 2025
Duration: 36 months
Workplace: Trento, Povo (Italy)
Benefits: flexi-time, company subsidized cafeteria or meal vouchers, internal car park, welcome office support for visa formalities, accommodation etc., supplementary pension and health fund, training courses, public transport, sports facilities, language courses fees. Further details at https://www.welfarefbk.info/
Application:
Interested candidates are requested to submit their application by completing the online form (https://jobs.fbk.eu/). Please make sure that your application contains the following attachments (in pdf format):
- detailed CV (including list of publications, statement of research interest and the name of two referees);
- cover Letter (explaining your motivation for this specific position);
- at least 1 reference letter.
Application deadline: Thursday, the 24th of October 2024.
Please read our Recruitment Regulations before completing your application.
For further information, please contact the Human Resources Services at jobs@fbk.eu.