Skip to main content

Platform overview

An outline of the core concepts that are helpful to understand before using Bitfount.

Projects​

Secure data collaborations happen in projects. Projects contain a collection of settings that configure the who, what, and how of the data collaboration taking place within it.

Projects are created and managed by a Project Owner who defines the terms and conditions and selects the machine learning or analysis task that will run. When collaborating with other users, Project Owners will invite participants to join, link their datasets and run the associated task.

Tasks​

A Bitfount task is the brain of a project. It specifies the algorithm(s) that will run on any dataset linked to the project which can include the use of AI models as well as other data science operations.

Tasks are defined and templated by data scientists via the Python SDK, making them available for repeatable use in one, or several projects. Once added to a project, the task configuration code can be inspected by any participating user.

Models​

Not all tasks include models, but generally, they are a key component of tasks. AI models are programs developed by data scientists designed to analyse datasets to find patterns and make predictions.

Off-the-shelf tasks provided by Bitfount host models owned by Bitfount, our partners, and even grant you access to use and test a whole library of open-source models. Data scientists can also integrate their own custom models to Bitfount and template them into tasks to be run in the app at the click of a button. Models can be public and open to any user, or private. Private models generally require a licensing agreement with the Project Owner before they can be used in a task that's used in a project.

Datasets​

A dataset is a collection of raw data records that can be used in Bitfount. Connecting a dataset to Bitfount is like registering it—only its metadata (name, description, and schema) is stored, never the raw data itself.

Datasets remain on the Data Custodian's (user that owns the dataset) systems at all times and are never transferred to, or stored by, Bitfount or any other collaborator's systems unless explicitly agreed. Once connected, datasets can be linked to projects, allowing tasks to run securely. Bitfount logs all tasks performed against the dataset, creating an audit trail for accountability and transparency.