Skip to main content


Welcome to Antigranular, your new workspace for privacy-first data science. Imagine it as the spirited cousin of platforms like Kaggle and DrivenData, a vibrant playground for you to engage in a friendly joust with data, whilst wearing the privacy armour.

A Community Platform

Antigranular is a community platform. Its goal is to give you the tools to build the "Eyes-Off Data Science" tools of tomorrow. Just like you, we also enjoy rolling up our sleeves, plunging into massive datasets, and learning new skills. However, what sets us apart is our commitment to privacy-preserving technology - in particular confidential computing and differential privacy.

Just picture this. You're scribbling away in your Jupyter notebook, dealing with a fascinating dataset. With our Python client and Jupyter extension, you can bring your script to life in a safe and secure environment. A simple %%ag magic command is all you need to transport your code block into a confidential computing space. The Python environment in the confidential compute is limited to ensure all sensitive data must go through a differential privacy mechanism, from one of a number of popular frameworks, before the results are safe to come back down to your local session.

Read more about:

Gamifying Privacy

Competitions? We've got those too. But we've given them a privacy-first twist. Success isn't just about acing the task, whether it's classification, regression, or something else entirely. It's also about how mindful you are with your privacy budget. Every bit of epsilon and delta differential privacy budget you spend costs you, so think twice before spying too closely at the data.

Through this blend of challenge and reward, we try to gamify "Eyes-Off Data Science". Here at Antigranular, we give you the reigns to navigate the utility-privacy trade-off, turning this crucial aspect of data science into a strategy game where everyone wins. So, we invite you to join the ranks of privacy pioneers on Antigranular, and be part of the exciting privacy revolution!

Read more: