DataRobot features a massively parallel modeling engine that can scale to hundreds or even thousands of powerful servers to explore, build, and tune machine learning models. Large datasets? Wide datasets? No problem. Modeling speed and scalability is limited only by the compute resources at DataRobot's disposal.
DataRobot employs all of the latest security protocols, including LDAP, Kerberos, and SSO/SAML 2.0 for user authentication to prevent unauthorized access to proprietary data. Data sent to be trained or scored is encrypted and transmitted with TLS 1.2 over your private network to the application or prediction endpoints.
Installation and model scoring in Hadoop
DataRobot can be installed as a service on YARN in Hadoop clusters, and perform distributed model scoring on data stored on HDFS, to bring automated machine learning to your Hadoop stack. Using Spark for large-scale data processing, DataRobot adheres to all Hadoop management processes and policies.
If your data is subject to strict regulatory compliance (e.g., HIPAA) or has a classification that requires specific security controls, DataRobot can be installed in an off-network configuration. Downloadable scoring code (or scoring code approximations) can be deployed to offline servers for local scoring.
Now that machine learning impacts an ever-increasing number of business processes, it is critical that a platform for building and deploying models is hardened, trusted and integrated with the ecosystem of technologies within an organization.
DataRobot is deployed by some of the largest Fortune 100 organizations around the world. Features that make DataRobot truly enterprise-grade include:
- “Four nines” of availability (99.99%)
- Distributed and self-healing architecture
- Seamless integration with enterprise security technologies
- Hadoop cluster plug-and-play
- Multiple popular database certifications