The development of autonomous maritime systems represents a complex technological challenge that demands sophisticated data integration strategies. This paper presents a comprehensive data platform that addresses the intricate requirements of maritime research and autonomous system development through a flexible, modular architecture optimized for heterogeneous maritime datasets. The platform integrates service-oriented and microservices architectures with Apache Pulsar messaging for low-latency data routing, PostgreSQL with PostGIS for metadata management, and distributed storage systems for multi-tier data lifecycle management. It supports synchronized multi-modal sensor data collection from uncrewed surface vessels, encompassing RGB cameras, thermal imaging, stereo vision, and LiDAR systems, demonstrated through operational deployment on the eM/S Salama testbed in the Finnish Archipelago Sea region. One of the most significant challenges in autonomous systems research is the limited availability of domain-specific machine learning datasets. The platform directly addresses this issue by allowing systematical collection of multimodal sensor data from marine environments while maintaining temporal synchronization, ensuring researchers have access to reliable data essential for developing artificial intelligence (AI) algorithmic models. The platform is designed to ensure compliance with European regulatory frameworks including GDPR and the Data Governance Act through privacy-by-design principles, implements Dublin Core metadata standards for interoperability, and includes Gaia-X compatibility for enhanced data sovereignty. The system is designed to handle concurrent data streams from multiple sensor modalities, addressing the lack of maritime-specific datasets for AI applications while enhancing data sovereignty and enabling collaborative research through a centralized integration hub that consolidates data from autonomous vessels, environmental monitoring stations, ...