top of page

We Provide an Easy-to-Use Platform for Hosting and Participating in Machine Learning Competitions

Screenshot 2023-11-20 at 12.20_edit

Our Unique

At, we understand that many organizations, especially in healthcare, are eager to utilize AI and develop models but face data confidentiality challenges due to regulations and sensitive user information. We address this issue by enabling organizations to create competitions and collaborate with our data scientist community. They can share their data without concerns about downloading, as our JupyterHub environment allows community members to access and train their models on the dataset without the ability to download it. This approach fosters open science and collaboration while safeguarding data confidentiality.


 Our Platform Features

Managed Jupyter Notebook

We provide ready-to-use Jupyter notebooks, which are customizable with CPU, GPU, and memory resources to suit competitive requirements. Participants can code in Python and R, with major machine learning libraries like Sklearn, XGBoost, PyTorch, and TensorFlow pre-installed and configured. maintains essential GPU drivers like CUDA, enabling users to simply open the notebook in any browser and begin coding and model training. The Jupyter notebooks come with persistent storage, allowing users to maintain their work over time.

Data Confidentiality and Security has invested significantly in securing its development environment. We enable organizations to host competitions with read-only datasets, meaning that participants can access the competition data exclusively through the JupyterHub environment without the ability to download it. However, they can still train their models locally. restricts outbound and inbound traffic to only those websites authorized by the organization.

Flexible Competition configuration provides versatile options for configuring a challenge. These include a built-in editor for descriptions that can include text and images. You can also embed a video to explain the competition's goals, add a logo, set an expiration date, and specify who can access your challenge. It can be made public for all users in the community or private, limited to a specific group of users, a corporation, or a particular region.

Automated Machine Learning

At, you can automatically generate machine learning models for regression and classification problems based on tabular data. You have the option to specify the compute time required to create your model. These models are developed using open-source libraries and ensemble methods crafted by Upon completion, you can download the model in Joblib format or run it on for generating predictions. Additionally, these models can serve as baselines for your ML competitions on, providing participants with an understanding of the base ML model they need to out perform.

Collaboration and open discussion provides a dedicated forum for each competition, facilitating communication and collaboration between the data science community and organizations during data challenges. This allows for easy exchange of technical questions, collaboration, and problem-solving among participants.

ML Competition Model deployment

With the platform, you have the option to deploy the competition-winning model on, customizing it with GPU/CPU resources as needed. You can then use this model to make predictions on new data through our user-friendly interface or via our API.

bottom of page