Topic
Data Infrastructure
Abstract
Design and implementation of a Proof of Concept (PoC) for a multi-tenant Retrieval-Augmented Generation (RAG) platform deployed on the ORFEO Kubernetes infrastructure. The project aims to provide a secure and scalable AI service for multimodal scientific data analysis, ensuring tenant-level data isolation while sharing LLM computing resources. Activities include the integration of authentication, API management, vector databases, and RAG orchestration components, the development of reusable deployment templates, and the validation of the platform on a real-world scientific dataset.
Host Institution
Scientific Supervisor
Tommaso Rodani tommaso.rodani@areasciencepark.it
Topic
Molecular Dynamics
Abstract
The project focuses on defining metadata standards, provenance capture, and FAIR-compliant workflows for simulation data, aligned with the emerging recommendations of the Molecular Dynamics Data Bank (MDDB). Activities include the analysis of existing GROMACS and PLUMED workflows, the design of reusable metadata schemas and RO-Crate packaging strategies, the development of prototype metadata extraction tools, and the production of a reusable DMP template to support future Molecular Dynamics as a Service within the ORFEO ecosystem.
Host Institution
Scientific Supervisor
Valerio Piomponi valerio.piomponi@areasciencepark.it
Topic
Marine Data
Abstract
Interpreting how vision and vision–language models represent continuous changes of their input requires stimuli whose generative parameters are known exactly. Existing controlled image collections rarely sample transformations along continuous, closed trajectories. They are seldom paired with text, and they lack the controls needed to separate structure built by a model from structure already present in the pixels. This thesis designs, curates and openly releases a dataset of rendered image sequences in which a single parameter varies along a closed trajectory, for several transformation classes applied to the same 3D objects: object rotation, camera orbit, hue, light direction, shape deformation and image translation. Each sequence comes with matched control conditions, per-frame ground truth, and text descriptions derived from the parameters, so the same transformation can be presented to image encoders, text encoders and vision–language models. The work covers a reproducible generation pipeline, a metadata schema (provenance, licensing, rendering settings) in Croissant format, automated quality control, and release under a persistent identifier with a held-out object split. The dataset’s utility is demonstrated by layer-wise geometric and topological analyses of a self-supervised vision encoder and a vision–language model.
Host Institution
Scientific Supervisor
Matteo Biagetti
matteo.biagetti@areasciencepark.it
Topic
Research Data Management
Abstract
The internship will take place at the Istituto Italiano di Tecnologia (IIT), an interdisciplinary research institute in Nanomaterials, Life Technologies, Robotics, and Computational Sciences, based in Genoa, in the heart of the beautiful Italian Riviera.
The Research Data Management (RDM) team, which brings together data stewards, software developers, and system engineers, is building new data infrastructure and services to support IIT researchers in managing their data throughout all phases of the research cycle. These efforts have resulted in an ecosystem that already includes services for data management planning (IIT FAIR Wizard) and for publishing and sharing research data (IIT Dataverse).
The aim of the internship is to contribute to the development of CORAL (COllaborative Research Application Layer), a research data platform based on iRODS (integrated Rule Oriented Data System). Coral supports the storage, organization, documentation, and discovery of IIT research data during the research data lifecycle. The intern will work with the RDM team to extend the CORAL platform with the new features informed by IIT researchers’ requirements. Activities may include identifying and developing use cases in collaboration with researchers to (i) automate data ingestion workflows, (ii) enable programmatic access to iRODS through the PRC (Python iRODS Client), or (iii) develop workflows for seamless data export from CORAL to community-recognized platforms (e.g., NOMAD), ensuring compatibility with their standards. Specific activities will be tailored to the intern’s background and interests.
The internship offers the opportunity to work within an interdisciplinary team and gain hands-on experience in implementing research data management strategies, ranging from understanding researcher’s needs and translating them into practical solutions to designing data infrastructure and workflows.
Host Institution
Scientific Supervisor
Valentina Pasquale
Additional Info
Reimbursement of up to 600 euros per month to help cover expenses (e.g., rent, transport, etc.) when working in person at our Genoa office.
Support with settling into life in Genoa.
Hybrid work mode. It can be arranged a mix of in-person work at our Genoa office and remote work, with a preference for in-person work at the start of the internship to help the intern settle in and get to know the team.
Topic
Cognitive Neuroscience
Abstract
Developmental neuroscience has long lagged behind other areas of neuroscience in adopting FAIR (Findable, Accessible, Interoperable, Reusable) data principles. Longitudinal study designs, heterogeneous measures, and small, hard-to-replace samples make open, well-structured data practices especially valuable yet rare in this field. This internship will explore how FAIR principles can be concretely implemented in developmental psychology and neuroscience research, with the goal of designing practical standards and workflows that support transparent, reusable, and shareable data. The timing is ideal: the SISSA BabyLab is currently being established from the ground up. The intern will join at this formative stage and will have the opportunity to directly contribute to building the lab's data infrastructure and analysis/sharing pipeline.
Activities:
Reviewing current standards and best practices for FAIR data management in developmental research
Identifying challenges specific to developmental data (e.g., consent, privacy, data heterogeneity)
Designing data organization structures and metadata standards for the SISSA BabyLab
Contributing to the development of reproducible analysis pipelines
Supporting the creation of data-sharing protocols compliant with ethical and legal requirements
Host Institution
Scientific Supervisor
Ambra Ferrari
Topic
Applied AI
Abstract
The project will be geared towards data-driven solutions by working with data analysis, predictive modeling, and applied artificial intelligence techniques. The students will collaborate with a multidisciplinary AI team on real-world projects, contributing to the design, evaluation, and improvement of advanced analytical and machine learning solutions. Practical training will enable the student to apply the principles and tools of data curation and FAIR-by-design management in an operational context.
Host Institution
Scientific Supervisor
Tomislav Subic tomislav.subic@arctur.si
Topic
Industrial Data & AI
Abstract
The internship focuses on managing and enhancing industrial data from manufacturing information systems and enterprise software. Activities may include developing pipelines for data extraction, cleaning and validation; documenting datasets and applying FAIR principles while respecting business confidentiality; developing machine learning models to identify patterns and anomalies; and exploring LLM-based assistants and Retrieval-Augmented Generation (RAG) architectures. The project will be tailored to the student’s background and interests, with opportunities to integrate the solutions into enterprise software services.
Host Institution
Scientific Supervisor
Angela Ingrasciotta, Marco Giacomini
angela.ingrasciotta@realcomm.it
marco.giacomini@realcomm.it
Topic
Marine Data
Abstract
This project aims to explore the Blue-Cloud data lake model and data curation pipeline, adapting them to the Gulf of Trieste observatory datasets. It will address challenges related to data harmonisation, metadata standardisation, quality control, and interoperability, enabling seamless integration of multidisciplinary long-term marine monitoring data and facilitating their analysis and reuse.
Host Institution
Scientific Supervisor
Alessandra Giorgetti agiorgetti@ogs.it
Topic
Materials Science
Abstract
The activity consists in the installation/customization of hardware components and in the development of software tools for the FAIRification of an electron beam lithography setup. The setup is composed by the EBL apparatus and a profilometer for the sample characterization.
The student will analyse the existing setup to identify any hardware and software shortcoming with regard to the FAIR data treatment and she/he will design and develop the missing components.
As a complementary activity, the student will outline a DMP for the instrumentation and will identify the proper (meta)data schemas also considering any eventual non-FAIR dataset produced by the laboratory, with the aim of harmonizing the laboratory output to the FAIR community standards.
Host Institution
Scientific Supervisor
Daniele Ercolani ercolani@iom.cnr.it
Topic
AI and Knowledge representations
Abstract
The objective of the internship is to analyse how the introduction of semantic models
(ontologies, taxonomies, controlled vocabularies, and knowledge graphs) can improve the
management of the company's information assets and provide the knowledge layer required for
the adoption of Agentic AI solutions.
The project aims to define a methodological framework and develop an application prototype
that integrates Master Data Management principles with modern Data Platforms.
Host Institution
Additional Info
Company-paid lunch for each day worked on-site
Abstract
Metabolic dysfunction-associated steatotic liver disease (MASLD) is a heterogeneous condition influenced by metabolic, inflammatory, and fibrotic processes. Numerous transcriptomic and proteomic datasets are publicly available; however, their integration is limited by differences in study design, patient characteristics, diagnostic criteria, disease staging, analytical platforms, data formats and metadata quality. This project aims to identify, curate and harmonize public transcriptomic and proteomic datasets from patients with MASLD and from cohorts with characteristics comparable to those studied at the Fondazione Italiana Fegato ETS, including obesity, metabolic dysfunction, type 2 diabetes, MASLD and liver fibrosis.
The student will systematically retrieve relevant datasets from public repositories and extract information on cohort characteristics, biological samples, analytical platforms, clinical variables and disease classifications. Particular attention will be given to the harmonization of diagnostic definitions, fibrosis and MASLD stages, demographic and metabolic variables, gene and protein identifiers, data normalization procedures and sample-level metadata.
Reproducible workflows will be developed in R or Python to standardize dataset structure, assess data quality, document missing or inconsistent information, and generate analysis-ready transcriptomic and proteomic matrices. The project will also define criteria for evaluating the comparability of public cohorts with the internal FIF raw datasets.
The expected outcome is a curated catalogue of MASLD-related public datasets, accompanied by harmonized metadata, data dictionaries and reproducible processing pipelines. This resource will support future cross-study analyses, biomarker discovery and the validation of molecular signatures associated with MASLD, fibrosis and metabolic disease progression.
Host Institution
Fondazione Italiana Fegato (FIF)
Scientific Supervisor
Pablo Giraudi
Topic
Green energy
Abstract
The project aims to analyse current research data management practices and design a flexible, standards-based architecture for experimental data generated by advanced battery, hydrogen, and electrification test facilities. Particular attention is given to semantic data modelling, integration with Piping and Instrumentation Diagrams (P&IDs), sensor lifecycle management, and the traceability, quality, and interoperability of data within dynamically evolving experimental environments.
Host Institution
Fondazione Bruno Kessler (FBK)
Scientific Supervisor
Michele Urbani, Luca Pratticò
murbani@fbk.eu, prattico@fbk.eu
Additional Info
Free Canteen when working onsite at Via Sommarive 18
Topic
Multi-Omics
Abstract
The project aims to define a preliminary institutional data governance framework, including recommendations for metadata, documentation, data organization, access, and sharing, aligned with FAIR principles. These guidelines will be validated through a pilot implementation within the Computational Biology Group by curating a subset of multi-omics datasets, harmonizing their metadata, and developing a prototype interface to enable dataset discovery and basic visualization.
Host Institution
Scientific Supervisor
Silvano Piazza, Marco Baralle
Abstract
What happens to a sensor reading before it informs a decision about a heating plant, a water leak or an energy bill? In this internship, you will work on that journey using real operational data from Area Science Park.
PARCO (Park Asset, Resource Consumption and Operations) is a dataset that combines more than ten years of measurements from over 200 sensors monitoring heating, cooling, electricity and water consumption along with tenant contracts, energy prices, weather records and forecasts. Already used in daily operations, its 15-minutes time series support daily tasks such as water leak detection, monitoring energy consumption and generation trough photovoltaic systems, cost allocation, periodic reporting and investment decisions. With this internship, you will join the Sustainability Office and work alongside maintenance staff and plant engineers to make the PARCO data pipeline more reliable, traceable and easier to maintain.
Host Institution
Scientific Supervisor
Fabio Morea
fabio.morea@areasciencepark.it
Abstract
This project focuses on the development of a FAIR-compliant data and service catalog for the Area Science Park data ecosystem, addressing the need for a multi-domain scientific metadata platform.
Throughout the project, key competencies are developed, including ontology navigation, pipeline development, production web application deployment on Kubernetes clusters, and identity provider integration.
Host Institution
Scientific Supervisor
Federica Bazzocchi