Quatra™
For executives desiring to launch AI competencies and practitioners transforming proof of concepts into enterprise AI, Quatra™ is a patent-pending data science platform that automates the AI lifecycle to ease the transition from business problem to scalable AI solution. Unlike piecemeal data science tools, Quatra automates everything: from feature engineering and algorithm selection to feedback collection and retraining. This enables anyone to uncover insights, regardless of technical expertise and frees teams to focus on business opportunities, not infrastructure and process. Plus, robust security and flexible deployment options ensure your data meets your governance policies and stays under your control.
HOW IT WORKS
Producers of data such as IoT devices, cameras, smartphones, and servers send data to Quatra. By using an automated lifecycle, the data science-as-a-service platform provides decisions and AI assisted responses to raw data, events and conditions. One-time configuration enables on-going automation of data preparation, feature engineering, algorithm selection, training, testing, data visualization, analytics, and learning from feedback.
BENEFITS
Fast Return on Investment
Simple deployment with time to insights in hours not days or months.
Low Total Cost of Ownership
Maintenance free AI that learns from feedback so a team of data experts is not required.
Democratize Data Analysis to Gain Competitive Advantages
Empower anyone to discover insights, regardless of technical expertise. Make data-driven decisions in real-time, outsmarting your competition.
ENTERPRISE SCALE AI PLATFORM
Data Console
AI Apps
Admin Console
On-Premises
KubernetesPaaS
Kubernetes Snowflake Redshift BigQuery Other OLAP
IaaS
AWS Azure Google Other
FEATURES
Simple Data Integration & Transformation
Connect, stream, and process data from diverse sources with ease. Connectors stream data from any source in real-time, from sensors to social media, databases to APIs. Our platform supports both data-in-motion and data-at-rest use cases, providing a unified and real-time data environment. Support for change data capture, event sourcing, and other patterns available. Choose from existing transformations or build custom transforms in Java.
Continuous Profiling & Feature Engineering
Continuous schema inference, profiling, and feature engineering enable effortless adaption to evolving data. Simply send data to the platform and the schema will be detected and data profiles provided to all components in the platform.
Bring Your Own Model
Tailor the platform to your unique needs. Deploy custom algorithms or choose from a repository that includes both community and platform-provided algorithms.
Continuous Learning & Feedback Loops
The platform continuously learns by repeating the lifecycle which includes data profiling, training, testing, ensemble execution, and more. Inputs and feedback to the cycle are raw data, end-user behavior, and administrator feedback. Embrace a dynamic environment where continuous improvement is the norm. Avoid concept drift and model decay!
Advanced Data Warehouse
Metrics are generated during stream processing and stored in a performant data warehouse, enabling swift analytical queries for actionable insights and data-driven decisions.
On-Premise, Cloud or Hybrid Deployments
A user account may be a member of one or more organizations. One cluster, which provides data processing services and stores data, may be used by one or more organizations. Each organization can use one or more clusters. Clusters have multiple deployment configurations such as on-premise, your desired cloud vendor, private cloud, or the multi-tenant cloud. The platform supports centralized and decentralized deployments of different cluster types.
Data Preparation
Automatic filtering of datasets based on entity relationships and ability to join other datasets. Anomaly detection and data quality monitoring available.
Intelligent Algorithm Lifecycle Management
Let our platform handle algorithm selection, training, testing, and deployment. Identify optimal algorithms, ensure accuracy and performance meet customizable thresholds, and store models in a centralized repository.
Ensemble Execution & Monitoring
Streamline complex workflows with ensemble grouping. Execute multiple algorithms within a group for every data record, providing comprehensive results. Monitor and fine-tune ensembles for peak performance. A weighted average is calculated from the results of multiple models to make a decision or prediction for the group. Choose from standard weighted averaging algorithms or bring your own.
Robust Governance & Security
Support fine-grained governance policies by defining roles that specify which actions are allowed or denied on what resources (e.g., allow creation of secured sharable link on stream with id stream-23).
Elastic Scalability
Whether you’re dealing with large volumes or velocity of data the microservice architecture of the platform scales elastically to meet throughput for real-time data science and AI.
Visualization & Analysis ConSole
Explanatory analytics and metrics of data processing are provided through a web application where users may drill-down, filter, visualize results, label data, and provide feedback.
UNLOCK PROVEN VALUE FAST
with our accelerated pilot program