Description du poste
About NAVER LABS Europe
NAVER LABS Europe is part of the R&D division of NAVER, Korea’s leading Internet portal and a global tech company with a range of services that include search, commerce, content, fintech, robotics and cloud.
The position
We define speech assistants as models capable of performing different tasks based on the user’s voice requests.
In this setting, it is important that speech assistants are not biased by the way users speak, or by their identity.
For instance, when different users give the same request, the system’s output should remain consistent.
One concrete example is the evaluation of spoken-dialect question answering, a setting where questions are asked by speakers with different accents.
We find this kind of evaluation relevant: a speech assistant that provides wrong or irrelevant answers depending on the speaker’s accent represents a serious deployment bias.
However, while speech assistants should not be biased by speaker characteristics, it may still be important for them to be aware of them.
For instance, we might be interested in a speech assistant able to consider the user’s expressed emotion, humor, or sarcasm when processing a request.
These tasks themselves may also introduce additional biases
- for example, along the gender dimension.
An assistant that systematically considers, for instance, women to be more emotive than men, is dangerously biased.
The goal of this internship is to extend an existing speech LLM to make it voice-aware, and to evaluate both its capabilities and the impact of the introduced tasks on bias.
The student is expected to:
Extend an existing speech LLM to be voice-aware.
The selected student will work on extending an in-house speech LLM that currently produces semantic embeddings from speech without preserving speaker information, making it voice-aware.
This includes designing and developing an acoustic block to be integrated into the existing implementation, and training new models.
Gather voice-centered datasets for bias benchmarking.
The selected student will survey the literature and create a collection of voice-centered datasets that are relevant to the task, and that will allow us to assess both the model’s capabilities and the potential bias introduced by these tasks.
Project context:
This internship is part of the Diké Collaborative Project, a French government-funded research initiative focusing on identifying and mitigating bias, fairness, and ethical issues in NLP model compression.
Funded by ANR (AAPG 2021, 2022–2025).
About the research team
In the Interactive Systems group, we develop AI capabilities that enable robots to interact safely with humans, other robots, and systems.
For a robot to be truly useful, it must represent its knowledge of the world, share what it learns, and interact with other agents, particularly humans.
Our research integrates expertise in human-robot interaction, natural language processing, speech, information retrieval, data management, and low-code/no-code programming to create AI components that empower next-generation robots to perform complex real-world tasks.
What we're looking for
Due to administrative restrictions, we can only accept applications from Master’s students for this internship
Knowledge of LLMs and speech processing
Strong programming skills
What we offer
We foster a collaborative environment dedicated to ambitious, multidisciplinary projects that translate advanced research into impactful, real-world solutions, supported by 30+ years of experience in AI and related fields.
Flexible work/life balance.
We are an equal opportunity employer that hires based on skills, experience, and merit.
We foster an inclusive and diverse workplace where all qualified candidates are considered fairly, regardless of background.
We’re based in Meylan, close to Grenoble, a city that offers the perfect balance of urban life, cutting-edge research and technology, and spectacular mountain landscapes that provide countless opportunities to relax, recharge, and enjoy the outdoors.
All applications will be carefully considered, even if not all required skills are met.
We value diverse backgrounds and the potential of each candidate, and we offer training to support the development of necessary skills.
NAVER LABS, co-located in Korea and France, is the organization dedicated to preparing NAVER’s future.
Scientists at NAVER LABS Europe are empowered to pursue long-term research problems that, if successful, can have significant impact and transform NAVER.
We take our ideas as far as research can to create the best technology of its kind.
Active participation in the academic community and collaborations with world-class public research groups are, among others, important ways to achieve these goals.
Teamwork, focus and persistence are important values for us.
When applying for this position online, please don't forget to upload your CV and cover letter.
Incomplete applications will not be considered.
NAVER LABS Europe is subject to French jurisdiction requiring organisations to stipulate that a job/internship is open to both women and men.
None of our jobs/internships are gender specific.
References
Beomseok Lee, Marcely Zanon Boito, Laurent Besacier, and Ioan Calapodescu.
2025.
NAVER LABS Europe Submission to the Instruction-following Track.
In Proceedings of the 22nd International Conference on Spoken Language Translation (IWSLT 2025), pages 186–200, Vienna, Austria.
Association for Computational Linguistics.
William Held, Yanzhe Zhang, Minzhi Li, Weiyan Shi, Michael J Ryan, and Diyi Yang.
2025.
Distilling an End-to-End Voice Assistant Without Instruction Training Data.
In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 7876–7891, Vienna, Austria.
Association for Computational Linguistics.
Feng, T., Lee, J., Xu, A., Lee, Y., Lertpetchpun, T., Shi, X., .., & Narayanan, S.
(2025).
Vox-Profile: A Speech Foundation Model Benchmark for Characterizing Diverse Speaker and Speech Traits.
arXiv preprint arXiv:2505.14648.
