Building a repository of data science and machine learning applications

As an increasingly large number of research projects produces machine learning models and data science applications a need arises for storing and sharing these research outputs, just like in case of e.g., produced research data. To meet this need, we built a custom repository focused specifically on sharing machine learning models and data science applications (SciLifeLab Serve, https://serve.scilifelab.se); it is currently available to life science researchers affiliated with Swedish research institutions. Our repository goes beyond just sharing files or code – users can make inferences using the submitted models and can interact with the submitted data science applications. At the same time, each submission is an entry with accompanying metadata. Our aim is to allow researchers to share their models and applications through an intuitive web interface and while doing so meet FAIR requirements. The service has components and, therefore, accompanying challenges of both a hosting platform and a research data repository. The developed platform is available as open source software that others are welcome to use to launch their own instances. In this poster we will share our journey so far, challenges, and how we met them.