- Product Architecture
Radically speed-up and enhance your AI development workflow. With live contextualized data, control of data quality, and a short path to production, moving from proof of concept to value takes hours, not months.
At every step of the data science workflow, Cognite Data Fusion (CDF) provides key tooling to help you succeed.
1. DATA DISCOVERY
With access to live contextualized data, you can move from searching to discovering. Access data through our SDKs, our Spark data source, or through our graphical interfaces.
2. PRE-PROCESSING AND FEATURE GENERATION
Speed up your pre-processing and feature generation by using pre-generated aggregates and our Spark data source.
3. LEVERAGE PHYSICAL SIMULATORS
Radically decrease the complexity of your modeling problems by leveraging physics simulators for feature generation. CDF is built to interface with several industry-standard simulators.
Use the programming language and modeling framework of your choice. CDF can also be used with all state-of-the-art autoML frameworks, including Google autoML, Datarobot, and H20.
5. DEPLOY AND SCHEDULE
Use Cognite Data Studio to rapidly put your model into production, or use any hosting environment of your choice.
6. QUALITY MONITORING
Set up quality monitoring of the data that your model depends on, get notifications, and trigger actions.
7. MODEL SCALING
Rapidly scale your model from one equipment or situation to thousands with contextualized data.
8. VISUALIZE, NOTIFY, AND COLLECT FEEDBACK
Rapidly provide your intended consumers with dashboards and visualizations, trigger notifications, and collect feedback.