Author: Božidara Pellegrini

This webinar is organized by the Slovak National Supercomputing Centre as part of the EuroCC project (National Competence Centre – NCC Slovakia) in cooperation with ThinkParQ.

The webinar will be held in English.

Abstract:

As high-performance computing and artificial intelligence workloads continue to grow in scale and complexity, efficient data access becomes a critical factor for overall system performance. Parallel file systems address this challenge by enabling concurrent access to distributed storage resources, supporting highly parallel and data-intensive applications.

This webinar introduces the fundamental concepts of parallel file systems and their role in modern HPC environments, with a focus on BeeGFS. The session will explain BeeGFS architecture and core components, key design principles, and features supporting performance, scalability, and ease of management. Topics will include data management capabilities introduced in BeeGFS 8, on-demand storage with BeeOND, and an overview of community and enterprise editions.

Real-world reference stories from HPC, AI, and enterprise environments will be discussed to illustrate how BeeGFS is applied in practice. The webinar will conclude with a discussion and Q&A session.

Speakers:

Dominic McKendry
Business Development Director Northern Europe, ThinkParQ GmbH

Dominic McKendry is Business Development Director for Northern Europe at ThinkParQ, the company behind the parallel file system BeeGFS. He works with partners and end users across the HPC, AI, and data-intensive computing landscape, focusing on real-world use cases and adoption of BeeGFS in research and industry.

He brings extensive experience from the enterprise IT and HPC sector, including over 18 years at Dell Technologies, and holds an MBA from Purdue University, with an academic background in international management and business studies.

Ingo Martini
Presales Director BeeGFS, ThinkParQ GmbH

Ingo Martini is a Presales Consultant at ThinkParQ, specializing in BeeGFS and its deployment in HPC, AI, and data-intensive environments. He supports customers and partners with solution design, architecture discussions, and technical positioning of BeeGFS.

He has a strong technical foundation built through long-term work in storage systems and infrastructure engineering, complemented by formal training and certifications, including AWS Certified Cloud Practitioner, and a solid engineering-oriented background.

Outline:a:

• Introduction and motivation: data challenges in HPC and AI
• Parallel file systems explained
• BeeGFS architecture and core components
• Why choose BeeGFS: performance, scalability, and ease of use
• Data management capabilities in BeeGFS 8
• BeeOND: on-demand storage for HPC workloads
• Community vs Enterprise (Hive) editions
• Reference stories from HPC, AI, and enterprise environments
• Discussion and Q&A

This webinar is organized by the Slovak National Supercomputing Centre as part of the EuroCC project (National Competence Centre – NCC Slovakia) in cooperation with NCC Portugal within the AI Accountability Dialogue Series.

The webinar will be held in English.

Abstract:

The EU AI Act introduces new roles and obligations that reshape how responsibility for AI systems is distributed inside organisations. In practice, however, AI ownership is often fragmented across legal, technical, compliance, data, and business functions, and further complicated by dependence on third-party and foundation models.

This webinar examines how organisations can address these challenges by distinguishing operational responsibility from operational ownership, and by clarifying decision rights and accountability across the AI system lifecycle. It discusses practical governance mechanisms aligned with organisational size and risk, including internal monitoring, documentation, and traceability of AI systems. Particular attention is given to common deployment challenges such as unclear ownership boundaries, reliance on external providers, and the emergence of informal or “shadow” AI use.

Speaker

Petra Fernandes

Lawyer – Data Protection, Artificial Intelligence & Cybersecurity

Petra Fernandes completed her Law Degree in 2003 and has since been advising clients on legal and governance matters related to data protection, artificial intelligence, and cybersecurity. She has served as a Data Protection Officer and as part of DPO teams for both private companies and public administrations.

In addition to advisory work, she regularly delivers training and awareness-raising sessions on data protection and AI governance for public and private sector organisations, with a strong focus on practical implementation and compliance.

Topics Include:

• AI ownership versus operational responsibility under the EU AI Act
• Roles and responsibilities of providers, deployers, and internal teams
• Proportional AI governance models for SMEs and large organisations
• Internal monitoring, documentation, and traceability of AI systems
• Managing ownership when using third-party and foundation models
• Addressing challenges such as shadow AI and informal AI use

Outline:

1. Introduction: Why AI ownership is more than a legal issue
2. The different players under the AI Act and their role in AI ownership
3. Provider and Deployer roles and internal organisational responsibility
4. Senior management accountability and decision-making authority
5. Proportional AI governance models
   • Internal monitoring and documentation
   • Mapping AI systems and use cases
   • Embedding responsibility into procurement and development
6. Challenges in real AI deployments
   • Fragmented ownership and unclear decision rights
   • Dependence on third-party and foundation models
   • Shadow AI and evolving systems
7. Key priorities for establishing clear AI ownership
8. Discussion and Q&A

Exploring Portugal’s Innovation Ecosystem

The first day of the mission was dedicated to visiting three key innovation hubs in Lisbon. The delegation began at AI Hub by Unicorn Factory Lisboa, a centre supporting startups developing solutions in the field of artificial intelligence. This was followed by a tour of Unicorn Factory Lisboa – Beato Innovation District, one of Europe’s largest technology campuses. In the afternoon, the delegation visited Taguspark, home to more than 160 technology and research companies. These visits provided Slovak participants with valuable opportunities for networking and deepening the technological dialogue.

Slovak–Portuguese Business Forum

The second day centred on the Slovak–Portuguese Business Forum “Green & Smart Futures”, held at Unicorn Factory Lisboa. The forum was opened by the Presidents of both countries, who emphasised the importance of “building bridges of innovation between the Atlantic and the heart of Europe.” The programme included presentations on the investment environment, contributions from SARIO and AICEP, signing of cooperation memoranda, and, above all, intensive B2B meetings during which companies identified technological synergies and discussed potential future collaborations.

Linking the Mission to the Role of NCC Slovakia

The participation of NCC Slovakia confirmed the growing interest of Slovak companies in solutions based on artificial intelligence, simulations, and work with large datasets – areas that naturally require high-performance computing resources.

Discussions during the B2B meetings showed that an increasing number of companies encounter infrastructure limits when developing AI models or processing large-scale data. NCC Slovakia helps them identify where HPC can deliver the greatest added value and guides them in designing projects that can effectively leverage advanced computational resources.

Within the EuroCC project, NCC Slovakia also provides educational support through free courses and expert webinars, and enables Slovak companies, universities, and research organisations to gain access to European supercomputing capacities via EuroHPC JU access calls. In this way, NCC Slovakia remains a key partner for Slovak innovation and research, helping transform technological ambitions into concrete projects powered by modern HPC resources.

This webinar is organized by the Slovak National Supercomputing Centre as part of the EuroCC project (National Competence Centre – NCC Slovakia) in cooperation with NCC Portugal , within the LLM Webinar Series connecting high-performance computing with artificial intelligence, culture, and innovation. The webinar will be held in English.

The webinar will be held in English.

Abstract:

Archaeological site detection is entering a new era thanks to advances in remote sensing and artificial intelligence. Archaeological sites such as hillforts often have irregular and complex shapes, making them difficult to identify using conventional computer-vision methods. Multimodal approaches that combine LiDAR-derived LRM images with aerial orthoimagery improve detection accuracy, but false positives remain a major challenge.

This presentation explores how Vision Transformers and LLM-inspired architectures can address these limitations. By using cross-modal attention mechanisms, these models integrate multiple data sources to enable precise boundary detection, reduced false positives, and scalable application across diverse landscapes and site types. A key element of this workflow is a human-in-the-loop refinement process, in which archaeologists review and provide feedback on model predictions. This iterative collaboration enriches the training data, enhances the model’s ability to distinguish true sites from background anomalies, and increases overall detection reliability.

Results from Northwest Iberia show a 99.3% reduction in false positives after a single refinement cycle, while nationwide deployment in England demonstrates robust performance across varied site morphologies. Combining multimodal fusion, transformer-based architectures, and expert-guided refinement, this approach delivers both accuracy and interpretability. The talk will conclude with insights into predictive modelling for identifying high-potential areas, accelerating large-scale archaeological surveys, and improving efficiency in heritage mapping.

Speaker:

Dr. Daniel Canedo – Research Fellow, Institute of Electronics and Informatics Engineering of Aveiro, University of Aveiro

Dr. Daniel Canedo received his Ph.D. in Computer Engineering from the University of Aveiro, Portugal, in 2024. Since 2017, he has been a Research Fellow with the Institute of Electronics and Informatics Engineering of Aveiro (IEETA). His research interests include computer vision and artificial intelligence, with a particular focus on their applications to complex pattern detection and image-based reasoning.

He has published in several international journals and conference proceedings and was awarded first place in the NATO StratCom Competition „How to detect malicious use of video and/or photographic content online” (December 2018, Riga, Latvia).

Topics Include:

• Vision Transformers and multimodal AI for archaeological mapping
• Combining LiDAR and aerial imagery for site detection
• Human-in-the-loop feedback for improved model accuracy
• Case studies: Burial mounds and hillforts in Northwest Iberia and England
• Reducing false positives through cross-modal learning
• Predictive modelling and future directions

Outline:

1. Introduction and Motivation
2. Vision Transformers: Extending the LLM Architecture to Image Processing
3. The Challenges of Detecting Archaeological Sites from Aerial Imagery
4. Use Case 1 – Burial Mounds: methodology, results, lessons learned
5. Use Case 2 – Hillforts: metodology, results, lessons learned
6. Conclusion and future directions
7. Discussion and Q&A

This webinar is organized by the Slovak National Supercomputing Centre as part of the EuroCC project (National Competence Centre – NCC Slovakia) in cooperation with NCC Portugal , within the LLM Webinar Series connecting high-performance computing with artificial intelligence, culture, and innovation. The webinar will be held in English.

The webinar will be held in English.

Abstract:

amália is a government-backed large language model focused on European Portuguese, designed to capture linguistic nuances, cultural context, and multimodal capability across text, speech, image, and video. The project aims for high-impact applications in public administration, research, and industry, with a strong emphasis on trust, alignment, and data sovereignty.

This presentation outlines the process and key methodologies employed by the amália LLM Team at NOVA University and Instituto Superior Técnico / UL, with a focus on utilizing European HPC resources such as the Marenostrum 5 (MN5) and Deucalion supercomputers. Prior Portuguese-language initiatives (like GlórIA) and European initiatives (such as EuroLLM) inform benchmarks and evaluation and help define the roadmap from beta to public release.

The core development pipeline is explored through three key dimensions:
1️ Initial data preparation and training, involving an extensive, multi-month process of transforming noisy HTML and PDFs into high-quality raw texts, followed by tokenization and core training methodologies in language modeling and instruction tuning.
2️. Model alignment, achieved through reinforcement learning approaches, Direct Preference Optimization (DPO) and Group Relative Policy Optimization (GRPO) with Verifiable Rewards (VR) to ensure safer and more trustworthy responses.
3️. Infrastructure setup for advanced RLVR (GRPO) training on MN5, which uses inference nodes for sampling and training nodes for collecting samples and running Verifiable Rewards, highlighting the complexity of configuring multiple custom environments (e.g., mathematics, programming, biology).

The talk concludes with key insights into the computational and methodological rigor required to efficiently develop state-of-the-art LLMs, positioning this work at the forefront of Europe’s innovation path in AI.

Speaker:

Prof. João Magalhães – CMU Portugal co-Director; Head of the Multimodal Systems Group, NOVA LINCS, Universidade NOVA de Lisboa

Prof. João Miguel da Costa Magalhães is a Full Professor in the Department of Informatics at Universidade NOVA de Lisboa and a senior researcher at NOVA LINCS. He serves as co-Director of the CMU Portugal Program and leads the Multimodal Systems Group. His research focuses on vision-language models, multimodal learning, and AI systems for semantic multimedia. He earned his PhD from Imperial College London and has been a key figure in Portugal’s AI and digital innovation ecosystem.

Topics Include:

• Building the amália model: architecture, training, and data curation
• Multimodal AI for text, speech, image, and video
• Cultural alignment and linguistic sovereignty in LLMs
• Evaluation, transparency, and responsible AI governance
• Future roadmap and collaboration opportunities

Outline:

1. Introduction: Why a Portuguese multimodal LLM
2. From Language to Multimodality: Scope and capabilities of amália
3. Data and Alignment: Linguistic diversity and cultural fidelity
4. Model Architecture and Training Process
5. Status and Roadmap: Beta achievements and next steps
6. Applications and Impact Scenarios
7. Discussion and Q&A