The competitions are the real fuel of the Crunch Foundation.

The data quality is top tier and clients offer big cash prize in order to crowdsource predictions.

There are for now different format of collaboration:

  • Limited in time competitions: The client buy the models to the rewarded winners.

  • Continuous competitions: last in time. The client pays to query the model through our API to receive the community's prediction each hour/day/week/months.

  • Rallies: Limited in time. It allows clients to test a dataset and the performance of the community models before launching a Continuous Competition.

  • Code Competitions: Participant must submit the code of their model in order for the customer to query inferences from the submission.

  • Predictions Competitions: Participant submit a simple prediction not the model or the code of the model.

Last updated