Service Catalog

Consultancy on opportunities to gain access to AI-oriented computing and data infrastructures available, notably but not limited to those at BSC. We analyse your technical needs and objectives, identify the most suitable access pathways, and guide you through eligibility, requirements and next steps.

Orientation service that helps new users understand how to access HPC/AI infrastructures, navigate user portals, and follow the essential start-up guides.

Advanced consultancy for organisations requiring in-depth technical guidance and coordination to access AI/HPC infrastructures, complementing standard L1 and L2 support processes.

Discover the right data for your AI project with expert guidance. This service helps you identify suitable external datasets and assess their relevance to your project, combining BSC-curated and partner data catalogues with expert support. It focuses on identification and preliminary assessment, and does not include data acquisition, preparation or validation.

Build reliable, FAIR-compliant and AI-ready datasets. Expert advisory support on metadata and documentation standards, compliance (GDPR, licensing, AI Act) and AI-readiness, so your data assets are reusable, compliant and ready to connect with other BSC AI Factory services.

Browse, explore and request access to BSC AI Factory datasets, with rich metadata across domains and data types. Controlled access to curated and governed datasets hosted or federated by BSC, without the need for acquisition, preparation or transfer.

Harmonise metadata, validate documentation and register outputs in BSC Dataverse with persistent identifiers, making your models and datasets visible, citable and ready for use in BSC AI Factory workflows.

Find out whether fine-tuning is worth it before you pay for it. We assess whether adapting a model is the right answer for your case, whether your data is ready for it (volume, quality, coverage, ownership and licensing, sensitivity), and what improvement and compute cost to expect, based on a reference demonstration rather than on estimates. A well-founded recommendation not to fine-tune is a normal and valuable outcome. The service focuses on assessment and guidance and does not include training your production model, deployment or operation in your environment.

Choose an AI model with evidence, not marketing. We define a benchmark protocol for your decision, measure the candidate models under identical conditions on tasks that look like your actual work, and report quality, cost and speed side by side —with their dispersion. You receive a reasoned recommendation, what would change it, and the protocol itself so you can repeat the comparison when a new model appears. The service focuses on evaluation and guidance and does not include software development, system integration or deployment.

Get better and more consistent results from generative AI, without hiring specialists. We compare established prompting patterns on tasks representative of your real work, measure which ones perform better, and hand you ready-to-apply guidelines with worked examples from your own context, along with the criteria to keep applying them on your own. The service focuses on guidance and evaluation and does not include software development, system integration or deployment.

Get answers based on your own documents, with sources you can check. We assess whether a retrieval-augmented approach fits your document collection and your questions, compare reference configurations on material representative of yours, and show how much each one adds over answering without retrieval. You receive an architecture recommendation with evidence behind it —including the option of not building it— before committing to an implementation.

Determining which legal obligations apply to you and whether your system falls under specific requirements of the European AI Act and related frameworks, together with an ethical self-assessment of the use case. Where necessary, the assessment is carried out with a legal specialist.

Identification of risks and priorities in fairness, robustness, privacy, explainability and transparency; prioritisation of what matters most in your context and actionable recommendations your team can keep applying on its own.

Expert consultancy for needs that do not fit any catalogue service, with joint assignment of the AI Development team and the relevant domain team.

Advice for hospitals and healthcare organisations, delivered jointly by the BSC Life Sciences domain, the bioinformatics team at the UMA’s SCBI and the AI Development team.

Analysis and advice on climate, environmental and Earth system use cases, delivered jointly by the BSC Earth Sciences domain and the AI Development team.

Expert sessions on language models and language technologies applied to your field.

In-person, guided training on AI and high-performance computing.

Self-study materials and modules available online.

Participation in events and joint outreach activities.

Access to incubation spaces and programmes within the ecosystem.

A first session to explain what the BSC AI Factory is and to understand where you stand: your activity, your processes and the problem you want to solve.

A sector-specific questionnaire that places your organisation’s maturity level and helps prioritise the next steps.

BSC AI Factory certificates attesting participation in the services.