Data Management

HammerHAI Data Management supports companies and research projects in organizing, documenting, and securely sharing their data for AI applications. The service combines private data inventories, controlled data access, and machine-actionable data management plans to improve transparency, traceability, and compliance throughout the AI lifecycle.

Target users

• Companies developing or deploying AI applications that require structured and secure dataset management
• Research and innovation projects looking to improve data documentation, sharing, and collaboration
• AI developers and data scientists seeking to track data provenance, document model training datasets, and address compliance requirements
• Project managers and data stewards responsible for data governance and data management planning

Description

Effective data management is essential for developing reliable, trustworthy, and compliant AI applications. Companies and research projects often work with datasets distributed across different systems, teams, and organizations. This can make it challenging to track available data, manage access permissions, and document how data is used in AI models.

HammerHAI supports companies in addressing these challenges by providing an integrated environment for AI-related data management. The service combines a user-friendly interface for managing data inventories with secure data sharing and structured data management planning.

The service builds on an extended version of the open-source data catalog platform Piveau, enhanced with access control capabilities. Unlike public data catalogs, the platform allows organizations to maintain private inventories of their datasets, making it possible to document and discover available data without exposing sensitive information.

For secure data exchange, the platform supports S3-compatible object storage with configurable access permissions. Datasets can be made publicly available or shared with authorized users through restricted access, supporting collaboration across organizational boundaries.

The platform also integrates with Data Stewardship Wizard, a tool for creating and maintaining structured data management plans (DMPs). This integration enables users to document data management practices through an intuitive interface while generating machine-actionable data management plans. These plans make data management information available for automated processing, validation, and integration into other workflows.

HammerHAI Data Management places a particular focus on the requirements of AI development. The service supports the documentation of data provenance, including the relationships between datasets, model training processes, and resulting AI models. Structured metadata and data management plans also provide a basis for answering questions about permitted data use, licensing, access restrictions, and compliance requirements.

The service offers the following capabilities:

Private data inventories

• Create and maintain structured catalogs of datasets within projects or organizations.
• Document datasets using standardized metadata while controlling who can discover and access information.
• Improve data discovery, reuse, and transparency without requiring public disclosure.

Secure data sharing

• Store and exchange datasets through S3-compatible object storage.
• Manage public and restricted access permissions for different datasets.
• Support controlled data sharing between project partners and collaborating organizations.

Data management planning

• Create and maintain data management plans using Data Stewardship Wizard.
• Use a guided interface to document data collection, storage, access, sharing, and reuse.
• Generate machine-actionable data management plans that can be processed by other tools and services.

AI provenance and compliance

• Document the origin and intended use of datasets involved in AI development.
• Track relationships between datasets, training processes, and AI models.
• Answer questions regarding data usage rights, licensing conditions, access restrictions, and compliance obligations.
• Improve the traceability and reproducibility of AI development workflows.

Applications

HammerHAI Data Management is particularly valuable for companies and collaborative projects that are developing AI solutions that involve multiple datasets, stakeholders, or regulatory requirements. Typical use cases include:

• AI project data management: Organize and document datasets across the AI lifecycle, from initial data collection to model development and evaluation.
• Secure project collaboration: Share datasets between companies, research institutions, and project partners while maintaining control over access permissions.
• AI provenance and traceability: Document which datasets were used to train or evaluate specific AI models, supporting reproducibility and accountability.
• Data governance and compliance: Maintain structured information about data ownership, licensing, permitted use, and access restrictions to support compliance assessments.
• Machine-actionable data management: Integrate structured data management plans into project workflows to reduce manual documentation and facilitate automated checks.

How to get access

The HammerHAI AI Data Management service assists organizations in establishing structured and secure data management workflows. The service includes support by the HammerHAI team.

Access to data.hammerhai.eu is provided as part of your user account. To apply for access, write to access@hammerhai.eu and describe your specific AI use case on HammerHAI.

Service provided by

Contact