ESRs Secondments at CNR, GPDP, University of Luxembourg and Intel

In this special edition of our blog posts, our ESRs write about their experiences and insights they made during their secondments.

Secondments enable ESRs to engage with prominent academics and practitioners at both partner universities and research institutes as well as industrial and regulatory entities. ESRs gain practical experience in how their research might translate into real-life problems encountered by businesses. At the same time, ESRs will be able to complement their research with practical experiences gained throughout the secondment and adapt their research projects accordingly. Finally, they also constitute an invaluable networking opportunity and grant our ESRs the possibility to identify possible career paths. Whether in the academic, regulatory, or industry sectors.

Each ESR will complete two secondments: one at a beneficiary (universities and research institutes) and one at the partners (industry or regulatory bodies) of the LeADS project.

ESR Soumia El Mestari at CNR and GPDP

As part of her PhD, Soumia spent around 5 months in Italy split into two secondments one took place at CNR Pisa from April to June and the second was in the Eternal City (Rome) at the Data Protection Authority GPDP.

During the first secondment, Soumia had the chance to exchange and collaborate with the CNR teams and ESRs that work on adjacent topics to hers. As Soumia’s research focus is around privacy issues in machine learning pipelines, her time in CNR was valuable since it allowed her to explore other horizons in the same topic as well as learn different approaches of privacy especially those related to federated learning and K-anonymity techniques. During that period, Soumia worked on different projects including the WOPA paper on fairness in machine learning systems and another collaboration project to study privacy attacks in federated learning settings.

In her second secondment in Rome at the GPDP Soumia got the chance to learn more about the legal aspect of data cnrprotection for machine learning systems. The exchange she had shaped the legal side of her thesis and equipped her with the necessary background to be as flexible in the legal discussion of privacy-preserving machine learning tools as she is on the technical side of the topic.

Both secondments were beneficial for the thesis and the ability to work with senior people and collaborate with them was extremely valuable. The secondment also added an aspect of discovering the hosting institution and its cultural surroundings. Being in Romeand Pisa was a pleasure and a trip into history that made the experience much more enriching. Interacting with people from different backgrounds made Soumia develop soft skills such as communication with law collaborators from different levels of expertise and exploring the viewpoints around the same privacy issues and how they can be influenced by the cultural backgrounds of each person.

ESR Louis Sahi at University of Luxembourg and Intel


Secondment at University of Luxembourg, April-June 2023:

sntThe main challenge of my research is evaluating trust and reliability in data processing using data quality criteria. This period at University of Luxembourg (UL) was an opportunity to discuss with several data management, cybersecurity, finance, entrepreneurship and innovation, and law teachers concerning a survey on data quality criteria. These exchanges allowed me to perform the methodology and results of my survey. Finally, they suggested that I interview data professionals from different domains to consolidate my academic findings. Then, we proposed a draft to conduct these interviews to get the best feedback from the data professionals.

Secondment at Intel, December 2023 – February 2024:

This secondment allows me to benefit from the expertise of Intel professionals and partners to consolidate my academic intelresults. As data quality is a current concern in the data processing ecosystem, I have provided a framework that lists 30 relevant data quality criteria in the literature review. From the definitions of the literature review, I proposed a unified and standardized definition for each criterion. At Intel Corporation, I had the opportunity to interview several data actors in the European area about data quality challenges. Then, I collected the opinions on the relevant criteria for each professional (their criteria, definitions and evaluation levels). Finally, I analyzed the feedback and compared it with my previous findings.