Leverage multiple controls to create an engaging data challenge. Participants can work individually or in groups, focusing on the same topic or tackling different ones.
With datasets mounted on instances pre-configured with all the classic ML/DL packages, participants can start exploring the data and building models immediately.
Bundle access to GPU instances for Deep Learning workloads and allocate credits at the group or individual participant level to ensure fair competition and controlled costs.
Use interim leaderboards to help participants gauge their performance and direction by comparing their results with those from other groups.
Combine model scoring and code quality analysis results to easily conduct comprehensive, multi-dimensional evaluations of submissions.
Gain direct access to the documentation and scripts to run the code and models again at your convenience, accelerating the internalisation of participants' work.
Full-fledged data science environment with VSCode and Jupyter.
Datasets instantly available as mounted volumes on the compute instances.
Metered access to GPUs with automatic shut down of inactive instances.
Easy sharing of intermediary data results within participating groups.
Kaggle-like automated model scoring with preloaded scikit-learn metrics.
Embedded leaderboards for positive emulation across participants.
HFactory allows you to make the access to the datasets conditional to the prior signature of an NDA by participants. You can set up the entire process directly on the platform, making it easy to track signatures and to download proof of signature documents.
Choose from various security levels and traceability options based on dataset confidentiality and participant profiles. Options include utilising a 100% European compute infrastructure, prohibiting local downloads of the datasets or using an internal GitLab instance with SSO.
Leveraging our expertise and network in Data Science and ML, we offer custom professional services to help you set up your own data competition.
Learn more about our process
ContactPrivacy