AI-ready Scientific Data Portal

Our client is a scientific data company helping pharmaceutical, biotech, and research organizations transform complex biomedical data into trusted, harmonized, and AI-ready assets. Combining scientific expertise with data engineering, bioinformatics, advanced analytics, and AI, the company curates and standardizes large-scale biomedical datasets, including single-cell, multi-omics, spatial, and clinical data. As a result, it makes the data easier to discover, analyze, and reuse across modern R&D and AI-driven research workflows.

The client's goal was to transform its growing collection of curated, harmonized, and AI-ready scientific datasets into an accessible digital product—a secure and scalable scientific data portal that researchers could efficiently explore and use.

With a large volume of complex datasets and rich scientific metadata, the key challenge was UX. It was important to provide an intuitive discovery experience, allowing users to quickly find relevant data through advanced filtering and clearly presented dataset and metadata information.

The other important issue was tight timing. The existing beta version had to be reviewed, stabilized, and made launch-ready ahead of a major industry event. Still, the client wanted to meet the deadline without compromising the platform’s future scalability and maintainability.

Our team joined the project to accelerate development and prepare the existing beta version for its initial industry launch.

The work included:

• Technical review of the existing beta to identify architectural gaps, usability issues, and potential show-stoppers before launch.
• Extending filtering capabilities to help scientific users quickly identify datasets relevant to their research criteria.
• UX improvements for the dataset catalog and metadata, focusing on presenting complex scientific information in a clear, structured, and efficient way.
• Technical stabilization and refinement of the existing solution to support the immediate launch while maintaining a scalable foundation for future new functionalities and data types. Currently, the portal only focuses on single-cell RNA sequencing (scRNA-seq), but the client plans to add other high-dimensional biological data types.

The product was successfully prepared for its initial launch at a leading industry event, where it received first-place recognition, providing strong early validation of the product concept and execution.
The client gained a stable and extensible platform for future product development, ready to onboard new scientific users and progressively expand its dataset catalog, filtering capabilities, and supported biological data types.

The technical and architectural improvements also established a foundation for long-term maintainability and ecosystem growth, supporting future integrations, new data sources, and AI-enabled scientific workflows.

• Backend API: Python 3.12, FastAPI, SQLModel ORM, Pytest
• Database: PostgreSQL, Alembic for database migrations
• Frontend: Angular 19, TypeScript, Node.js 22
• Authentication: Keycloak
• Infrastructure: Docker containers